For the love of baby unicorns: My Code4Lib 2018 Keynote

I was honored to give a keynote at the 2018 Code4Lib conference on February 14. The livestream is already available (keynote starts at 1:26, and the conference hashtag was hopping.

First, a confession, I am not a big fan of keynotes in general — the whole idea that one person would stand in front of the group and spew some special wisdom and/or inspiration to the crowd is a little too sage on the stage for me. And I feel especially self-conscious in front of this crowd, since I have never coded for libraries. In fact, the last time I wrote anything resembling code was 15 years ago and it was SPSS code to run statistical analyses on my dissertation data. I’ve never coded for libraries, and I’m pretty sure my current career trajectory means I never will … and we can probably all be grateful for that.

So, I’m not a big fan of the whole keynote thing and I have virtually no expertise or experience in the kind of work most of you do. But, the fact that you all pick your keynote speakers via community voting tells me that some critical mass of code4libbers must want to hear something from me. Plus, I figure we have to support democracy wherever we find it these days.

But I’ve struggled a bit about what I could talk about it — my approach to keynotes is to try to figure what I might be able to bring to the specific audience that they might not get from each other or from some other speaker they might have selected.

But ever since I got to MIT, that means folks expect me to explain gravitational waves or quantum physics of artificial intelligence. And I’m willing to fake it in a room full of my fellow directors, but I won’t try that here.

Or maybe you want to hear me tell you about how MIT Libraries is going to build an “open global platform” of abundant, equitable, interactive access to information — but to be honest, I can only talk about it as a metaphor.

Its my colleagues Heather Yager, Helen Bailey, Matt Bernhardt, Osman Din, Mike Graves, Jeremy Prevost, Richard Rodgers, Andromeda Yelton and Solh Zendeh are actually going to do the building and can talk about it. Heather is MIT’s AD for Digital Library Services, and the rest of those awesome folks are our Digital Library Engineering Team — oh and by the way, we are hiring a Department Head for that team, and I’m especially keen on increasing the number of people of color and white women in library tech, so if you or someone you know might be interested, let me know.

Anyway, I do have some things to say, because what I came up with is that I do bring to librarianship a sociological lens and a feminist perspective, and I have done a few talks on bias in technology and lack of diversity and equity in libraries.  And it is tempting to give a ranty ALL CAPS keynote about all the problems of lack of diversity in tech and lack of progress on creating welcoming and inclusive cultures, but I’ve done that; and plenty of other people have as well.

Plus, I just came from the ALA midwinter conference, and my friend Pulitzer Prize winning novelist Junot Diaz, who was one of the headline speakers, did the ranty thing way better than I could.

And girl howdy, he did not hold back. For example – he implored us to recognize that “A profession that is 88% white means 5000% agony for people of color, no matter how liberal and enlightened you think you are.” He also said we need to have “a fucking reckoning” about the pain we cause, and that we need to do some hard work on decolonizing our organizations and our professions. And y’all, he ain’t wrong, and what goes for libraries, goes for library tech too. But I figure y’all already know that, or if you don’t then another liberal white lady, even an obviously queer one like me, preaching repentance one more time is not going to make much difference.

I feel like we have been talking for a very long time about lack of diversity in tech, and in library tech, and we’ve critiqued the culture and we’ve adopted codes of conduct (yay), and we do better at getting diversity in our speaker line-ups …  progress has been made, but not nearly enough; not on gender, not on race, not on creating truly inclusive cultures in library tech where members of marginalized groups are recruited, retained, mentored, promoted, and genuinely provided with equitable opportunities to contribute, thrive and lead.

And from what I understand, code4lib is actually a community that does many things right. In this community, people feel included, they feel safe/r, sexism, racism, transphobia, ableism and the like are not tolerated. In fact, I sense that many folks from marginalized groups consider the code4lib community and conference to be a kind of haven. And that’s great, except that it isn’t ok that folks need a haven from their own work organizations.

So I thought instead of a ranty keynote with lots of examples of the problems, I would share some solutions — some from recent research I’ve been reading, some are based on stories I’ve heard, and some of what I’ll say is just me trying to use the privilege I have as a library director to say some things that others maybe don’t feel as safe saying, or don’t have access to the kind of platforms I do for saying it.

I’m going to start with the assumption that we all agree that lack of diversity is a problem, and that we want cultures that are welcoming and inclusive.

And I should also throw in the caveat that there will be things I say that are primarily meant for the white guys in the room. For some of you, this may feel a bit like preaching to the choir, but I hope instead that it feels like equipping the choir.

Also, I’m focusing primarily on inclusive cultures, because as much as we want to throw our hands up and claim diversity is a pipeline problem, the retention data tells us that we have problems with toxic work cultures and unfair practices driving women and/or people of color and/or LGBTQ folks out of tech as well. I highly recommend a report I just read called the Tech Leavers Study, put out by the Kapor Center for Social Impact. The report contains some really sobering stats on the degree to which unfair treatment, discrimination, stereotyping and bullying in the workplace are the reasons why people from marginalized groups leave tech work:

• among people who leave tech jobs, unfairness or mistreatment is the leading reason; people are 2x more likely to leave because of unfairness or mistreatment than because they have been recruited by a better opportunity
• women of all backgrounds experience & observed significantly more unfair treatment than men
• LGBTQ employees were most likely to report bullying, and to experience public humiliation and embarrassment
• In short, unfairness, discrimination, and crappy treatment drives turnover; and it suck more for underrepresented people of color and LGBTQ folks
• AND 57% of folks who left tech would have stayed if their prior employer had addressed workplace environment and had created a more fair and inclusive culture.

So lets dig in and talk about things we can actually do to improve those cultures, and move the needle on diversity and inclusion.

There are individual things we can do, workgroup things we can do, and leadership things.

At the individual level, it is actually not enough to simply not be a blatantly racist, sexist, homophobic jerk. That’s a start, but it is not enough.

The more subtle forms of stereotyping, discrimination, and harassment take their toll, and are shown to lead to lower retention in tech jobs for marginalized folks.

One thing men can do, and white people can do in mixed race interactions, is stop mansplaining, and stop whitesplaining. Just stop.

Stop doing it in person, and stop doing it online. What I mean is simply this — if you have some power and privilege in a situation (and if you are white and/or a dude, chances are you do have some privilege; if you are a white dude, you definitely have power and privilege), then for the love of baby unicorns please refrain from giving your unsolicited advice and opinions to others. Practice some restraint. Just try to sit on your hands for a minute and entertain the idea that maybe you don’t have to jump in.

Don’t be the guy who reads the blurb for Safiya Nobles’ new book Algorithms of Oppression and decides on the basis of a couple of hasty google searches that her research is faulty, that he has proven it, and that he needs to inform twitter.

And don’t be the guy who responded to the article Bess Sadler and I wrote about Feminism and the Future of Library Discovery a few years ago by tweeting that only people who don’t know anything about building search tools would write an article like that. Those of you who know Bess know that she knows a thing or two about building search tools, and has spent a minute or two actually doing exactly that.

I’m sure it often feels like other people really need to hear what you have to say about a topic, but if you are a white guy I’m going to ask that you consider all the evidence we have about how women and/or people of color are talked over, ignored, and shut out of conversations; and think about the fact that the airtime you take up may be the airtime that could have gone to a woman and/or a person of color; and maybe you could just decide you don’t have to weigh in on everything. and you definitely don’t have to explain things to women and/or other marginalized peoples.

But what you can do, and research shows that this works to break down stereotypes, is to publicly vouch for the people you work with who are members of a marginalized group. This is something anyone in leadership or other positions of power can and should do, and it is something men can do for women colleagues and something white folks can do for colleagues of color. Vouching for someone means you talk up their accomplishments and skills, especially in a group setting that is predominantly male and/or white. Use your power and privilege to vouch for others. It is the opposite of mansplaining.

Vouching for someone might look like this “Hey Tom, have you met my colleague Safiya? She is an expert on algorithmic bias in search engines, and has just published a book based on years of research” or “This is Bess, she has tons of experience in developing open source search tools, and was instrumental in the initial development of Blacklight.”

Another thing individuals can do to improve the culture and make tech more inclusive and welcoming is to be an active bystander and ally. If you see something, say something. If other men are talking over women, jump in and say “Hold on dude, I really want to hear what Cathy was saying.” If a white person repeats something a person of color said and is getting credit for it, say “That sounds a lot like the idea Bergis suggested. I would really like to hear more about it from Bergis.”

There are some things we can do in our workgroups too.

Just like all politics is local, all culture is local.

Or at least, lots of the visible manifestations of culture are local. There is research that shows that workplaces that are plastered with stereotypically “tech or nerd guy” cultural images – think Star Trek – have negative impact on women’s likelihood of pursuing tech work and of staying in tech work in general or in that particular work environment. Replace the Star Trek posters with travel posters, don’t name your projects or your printers or your domains after only male figures from Greek mythology, and just generally avoid geek references and inside nerd jokes.  Those kinds of things reinforces the stereotypes about who does tech; and that stereotype is the male nerd stereotype.

I also want to urge you all to pay attention to the kinds of informal socializing you do at work and in those liminal spaces that are work/social – if all the guys go to lunch together and not the women; then maybe stop doing that. And if the guys go to lunch and talk about women, then really, really, really stop doing that.

If there’s a core group of guys who go out for beers after work just because you’re all friends, that’s kind of OK; but if you also talk about work and make decisions then it is definitely not OK.

Be more aware. Be accountable to each other for being inclusive. And if you keep inviting people and they keep saying no, don’t keep inviting them and expecting them to suddenly say yes, and don’t shrug it off and say you tried; try something else instead. Try asking whether there is something else more folks would want to do. Consider that there might be a better time for informal team building that would work for everyone. And if the idea of giving up your beer time with your buddies seems like a bridge too far … then at least admit that might mean that you value your all-dude happy hour more than you value an inclusive work culture.

So now that I’ve told you to give up your Star Wars posters and your bro-time; the rest of this – the organizational level suggestions — are going to seem easy.

At the organizational level, a recent report from the Kapor Center for Social Impact called the Tech Leavers Study found that individual diversity and inclusion initiatives are not nearly as effective at increasing the retention of women and/or people of color and/or LGBTQ folks in tech as a comprehensive, integrated approach.

A comprehensive approach includes:

• Top level commitment to D&I, including stated values and goals, and an approach that treats diversity as a strength – leadership needs to truly understand that diversity isn’t just nice to have, diverse teams are better, more effective teams.
• Training in a range of topics, including unconscious bias training, bystander intervention, understanding micro-aggressions, preventing sexual harassment, how to be a trans ally, etc.
• A commitment to gathering data, and looking at the data. All kinds of data — Hiring data, retention data, internal promotion data, and survey data designed to uncover feelings of inclusive or exclusion across demographic groups.
• Developing and maintaining confidential channels of communication so that issues of bias and discrimination can be expressed and handled.
• Commiting to auditing management practices – hiring, promotion, compensation, and internal work assignments.

 

One of the really, really important findings about diversity initiatives is that training alone is not very effective. In fact, in some circumstances, unconscious bias training alone can backfire – because white folks get resentful, and marginalized folks are put in the spotlight and scrutinized.

OK, so I’ve made lot of suggestions, and I didn’t put any of it on slides; so let me summarize, plus I promise to send the text, with links to some of the studies I cited by end of the week.

So to sum up, if you really want more inclusive library tech:

• Instead of mansplaining, vouch for women and people of color, and use your privilege to intervene when you witness micro aggressions and silencing
• Pay attention to small group culture – ditch the Trekkie posters and such, and be intentionally inclusive in social activities
• Take a comprehensive approach to diversity & inclusion that goes beyond 1 shot training and includes clear leadership, data, accountability, and communication channels
• And I’ll suggest, monitor progress on diversity and inclusion the way you would monitor progress on any technical system or technology project. Check progress regularly, pay attention to data, and be transparent.

I’m actually really excited about the potential for new library technologies and technical approaches that will empower users, open up access to more collections, and allow for transformative uses of library content. But if we don’t make a commitment to progress on our diversity and inclusion challenges, then I really believe that the tools and technologies we create will not be as good or as useful to the range of users and uses we should be serving.

I dislike traditional Q&A’s almost as much as I dislike traditional keynotes. So if anything I said inspires a thoughtful comment, go ahead – I don’t need to stand up here and be the person with all the answers. And if you want to ask a question entirely unrelated to what I talked about, go for it.

The only thing I ask is that we make sure that women, people of color, all the queers, and other marginalized folks who want to speak get the first shot at the mic.

 

What happens to libraries and librarians when machines can read all the books?

Revised text of talk I gave for the Harvard Library Leadership in a Digital Age program.

The description of this course promises that “you will identify fundamental changes occurring in the field of knowledge management and consider their implications for libraries, information services, and library leadership.”

I think my session maybe breaks the rules a bit (which is my first leadership tip for you: when it feels like the right thing to do, break the damn rules!).

One of the things I think is important for library leaders is that we look at fundamental changes outside of knowledge management and consider their implications for libraries and the work we do.

I think looking outside of changes in our own field is essential if we want to be active, effective leaders who don’t merely respond to change, but who create and shape the change we believe is needed in libraries and archives.

So, I want to talk about AI and libraries in at least 2 ways:

1. Substantively, I want to share with you some of my thoughts and speculations about the potential implications of AI and machine learning for libraries and librarianship .
2. I also want to talk a bit about AI on a more meta-level – that is to say, I want to use my own commitment to learning about and thinking about AI and its implications for libraries as an example of the more general tension leaders face between tending to immediate, local challenges and thinking about, preparing for and creating the future.

So let’s start with why I’m interested in machine learning and AI.

Basically, it is because I think that it is past time for us to take digital libraries to the next level; and I think the next level is likely to involve machine learning and optimizing our collections, services, and spaces for machine learning applications.

Where are we in digital libraries right now? We are still in the midst of the initial digital shift in libraries (really from the mid-to-late 1990s to now).

In this shift, we have gone from libraries being a place where individuals came to find physical books and journal articles (and manuscripts, and images, and lots of other stuff) so that they could read those books and articles themselves, to libraries being a service that individuals use to gain online access to journal articles, and e-books, and digital images and manuscripts and more so that they can read and use those things on their own digital devices.

This ongoing digital evolution of libraries and of how students and faculty use scholarly content is significant and has arguably made research and teaching more efficient and more productive.  The advent of online and digital libraries has also made more information more accessible to more people than ever could have been possible when scholarly materials were available only in tangible, physical formats.

But if this switch, from individuals reading books and articles one at a time in print to individuals reading books and articles one at a time on their own digital device is all we get from the digital revolution, then it won’t have been much of a revolution.

In the title of the talk, I ask “what happens to libraries & librarians when machines can read all the books?” But the truth is that we are already there – or at least the machines are. So it behooves us to be ready for it – intellectually, strategically, and operationally.

I think an important part of leadership is not just responding to changes, but actually getting in front of those changes when we can.

Let’s start with some definitions.

From the MIT Press Essential Knowledge book Machine Learning:

AI is “Programming computers to do things, which, if done by humans, would be said to require “intelligence”.

Machine learning is a kind of AI, where the computer is programmed to optimize a performance criteria using examples or past experience. The machine does what the data tell it to, not what a program tells it to.

In describing the advent of machine learning, Ethem Alpaydin says:

“nowadays, more and more, we see computer programs that learn – that is software that can adapt their behavior automatically to better match the requirements of the task. We now have programs that learn to recognize people from their faces, understand speech, drive a car, or recommend which movie to watch … once it used to be that the programmer who defined what the computer had to do … now, for some tasks, we do not write programs but collect data”

Since I’m not a computer scientist or an engineer, I use the terms in relatively loose ways and often interchangeably.

When I think of AI and machine learning in the context of libraries, I think of computer programs and algorithms that can extract/derive meaning and patterns from data, make predictions and inferences about and with new data, and in doing so, solve problems at scales not possible by humans only

At an MIT symposium a few years ago Elon Musk, CEO of Tesla, talked about the existential threat of AI and suggested a need for regulatory oversight. Specifically, he said “With artificial intelligence, we are summoning the demon.”

So let’s talk about the fears and concerns, maybe they aren’t as existential as Musk’s (I find librarians tend to be more practical), but I’m sure we have some. I certainly do.A

What are the dangers of AI, especially in relation to libraries and the things we support — especially research & learning? Here are some common concerns:

• Robots will take our jobs – In an article in Library Journal in April 2016, Steven Bell writes about the Promise and Peril of AI for Academic Librarians – and he asks “Could artificially intelligent machines eliminate library jobs?
• One reason people argue that AI will not replace library or other jobs is that machines can’t replace the deeply human skills of creativity and interaction; which may mean that those skills become more valuable or could mean that AI will usher in an era where creativity and empathy are devalued and rare
• Another fear is that AI will eliminate the relationships between people and books, and between librarians and their community members
• And one concern I think is very important to take seriously is the reality that without explicit counter-measures, machine learning & AI will re-inscribe & magnify existing systems of inequality and racism, sexism, homophobia and the like.

Here’s a cautionary tale about that last concern:

Last Spring, Microsoft unveiled a twitter bot named Tay; programmed to tweet like a teen. What could go wrong, right?

Tay is backed by Artificial Intelligence algorithms that were supposed to help the bot learn how to converse naturally on twitter. But what happened is that the bot learned quickly from the worst racist sexist corners of twitter – and within 24 hours Microsoft had to shut the experiment down because the bot had started tweeting all kinds of sexist, racist, homophobic, anti-Semitic garbage.

Even, or especially, with those concerns in mind, I think we need to think about AI and machine learning and the implications for libraries.

My thinking about AI, machine learning, & libraries, is guided by 3 kinds of questions:

1. What role can libraries play in making sure we don’t summon the demon; or at least that we have the tools to control or tame the demon?

2. How might we leverage AI in support of our missions? How might AI help us do some of our work better?
3. How might we support AI and machine learning in ways that are consistent with and natural evolutions of the long-standing missions and functions of libraries as sources of information and the tools, resources, expertise to use that information?

Let me address the 1^st issue and offer some thoughts on libraries as demon-slayers in our AI future. First, we need to accept that AI and machine learning are becoming more prevalent in our daily lives, and in many learning and research contexts.

Then we have to think about what concerns around AI that libraries and librarians are maybe especially well-suited to addressing; like privacy, context, authority, and ensuring the data used to train AI is inclusive and diverse and of high quality.

This last one seems to be to be especially urgent – as an example, when Apple hired a new Director of AI research, he spoke about the promise of AI as a research tool, imagining — “If I ask you something about a particular thing, can your system basically go to Wikipedia, read a few different articles, learn some facts about the world, and provide you with the right answer?” As much as we all love and use Wikipedia, I suspect that makes some of us cringe. Wouldn’t it be better to have “your system” go to the actual scholarly literature on a topic?

The 2^nd area we should think about is how we can leverage AI in our work?

A typical area we think about is reference – this is Steven Bell’s concern that AI chat bots will replace reference librarians.

There is also plenty of potential around using machine learning in search – the 2 articles that were assigned reading for this session cover that ground fairly well (see list of references at end of this post).

We might also imagine leveraging AI for recommendation systems, and for cataloging and organizing our collections.

What if we turned my original question around and asked what would we do if librarians we could read all the books?

If we really could absorb all the information in our collections and make some sense of it, what would we do? What could we do if we had the capacity to read all our books, and maybe all the books in our peer libraries, and derive patterns from them?

What would we do that we can’t do now? What would we do better that we already do?

Can thinking about AI and machine learning in that way help us conceive of ways to leverage the fact that machines actually can do that now?

Finally let’s talk about how machine learning and AI might change or be changing research; and how we might start to think about optimizing our libraries to support new kinds of research made possible by text & data mining, AI and machine learning.

Let me share 2 really interesting examples:

Prof Regina Barzalay and her students and colleagues at MIT are using machine learning to extract information and predictions from the unstructured data in tens of thousands of pathology reports. Faster, as accurate as humans; and based on much larger amount of data than humans have access to.

Another example I learned about from my colleague Sara Lester, Engineering Librarian at Stanford, is GeoDeep Dive is a tool for geologists that uses machine learning to extract data that is buried in the text, tables, and figures of journal articles and web sites, sometimes called dark data, about rock formations.

GeoDeepDive is based on open source code, that can be repurposed on other datasources. Should libraries be exploring how tools like this could help us extract even more meaning and information from deep within our collections?

I think it is important not just that we know about these kinds of efforts, but that we proactively ask where can AI and machine learning be leveraged in the service of better science?

And how do libraries leverage our resources and skills to ensure it really works – and is infused with and informed by values we care about (inclusion, privacy, democracy, social justice, authority, etc.)?

Where can we intervene to make sure the research based on AI and machine learning is as good as it can be?

We help students find the best books and articles for their learning; so can we help programmers find the best data for their algorithms to learn on?

Can we help them think about the questions they want their machine learning applications to answer? Can we help fit the data to the question?

A final string of thoughts, provocations, and questions that keep me up at night:

As I begin to fully appreciate the fact that machines really can read all the books, and can “learn” from them; I am convinced that we need to think more rigorously about reading.

What are the different ways of reading, and what are the various goals of reading?

What can we learn best, as individuals and/or as society, through human reading? what can we learn best through machine reading?

Can we start thinking about how to design libraries to maximize the unique payoffs of many different kinds of reading?

How can texts (and images, and data) be maximized for human discovery and reading? For discovery via algorithms and reading by machine learning applications?

What does it mean to maximize our collections for humans and what does it mean to maximize them for machines and algorithms?

OK – really wrapping it up now:

Machines can already read all the books. Or at least they can read all the books (or articles) that they can read.

(sidebar about how the proliferation of AI should compel us to double-down on mass digitization and on open access)

Trying to understand a little bit about AI and machine learning has taken me way outside my cognitive comfort zone, but I think it is the kind of thinking we need to do to be effective library leaders and to be effective stewards of the future of libraries, librarianship, and for those of us in research libraries, for the future of scholarship.

I think it will be crucial that we avoid the temptation to continue to serve primarily individual human readers and let the computer scientists worry about how to apply machine learning and AI to vast libraries of resources.

I think we would be wise to start thinking now about machines and algorithms as a new kind of patron  — a patron that doesn’t replace human patrons, but has some different needs and might require a different set of skills and a different way of thinking about how our resources could be used.

For further reading:

• “New AI-Based Search Engines are a “Game Changer” for Science Research”, by Nicola Jones, Nature magazine on November 12, 2016
• Searching for Lost Knowledge in the Age of Intelligent Machines, by Adrienne Lafrance, The Atlantic. (audio version also available) 

• Machine Learning: The new AI  by Ethem Alpaydin (part of the Essential Knowledge Series from MIT Press)
• And more

Serendipity as prick

(How’s that for a click-baity headline? Am I doing it right?)

This is just some fleshed out notes from a short talk I gave to kick off the Codex Hackathon at MIT this weekend. My instructions were to get folks excited about the theme of Serendipity (not coincidentally, a favorite topic of mine).

I’m obsessed with serendipity.

Serendipity gets a bad rap sometimes, and is often associated with some pretty sexist and ageist stereotypes about bun-headed, pearl-clutching librarians who cling to their books and their browsing just in case someone might discover some unfinished, never before played sonata by German Composer Robert Shummann; or some unknown, uncatalogued, forgotten short stories by Zora Neale Hurston.

Those kinds of things actually do happen — those actual examples happened.

But Serendipity – at least the kind I am obsessed with is about so much more than just stumbling on some unexpected treasure in the bowels of libraries or archives … Not that there’s anything wrong with that.

I hope we can all think expansively and creatively about serendipity – about what it is, why it ought to be encouraged and facilitated (especially now), and how we might leverage technology to democratize access to serendipitous encounters.

The OED says Serendipity is a happy, unexpected, accidental encounter with new information (paywall, so no link. But trust me).

I think of Serendipity as exposure to ideas, people, perspectives that we didn’t know we wanted to or needed to be exposed to.

And lately, I’ve been thinking of Serendipity as a pin prick that might burst our filter bubbles.

The most famous quote about serendipity is attributed to Louis Pasteur, who is said to have said “Serendipity favors the prepared mind” — which if you ask me is a little elitist.

In a print-based world, scholarly serendipity certainly favors a well-connected and well-placed mind.  In the print-based world; happy, unexpected, accidental discoveries of ideas, and knowledge and perspectives is only possible for those with access to the people and the books and libraries and archives where those ideas reside. And that access has always been limited to elite communities.

The internet and the growth of digital libraries holds the promise of democratizing access not just to knowledge, but to the opportunities to discover things you didn’t know you wanted to discover.

In my ideal world, the likelihood  of a serendipitous discovery is limited only by our openness to the possibility.

We need serendipity now more than ever – and we need it for as many people as possible. Because encountering new, unexpected ideas and information – being exposed to data, arguments, concepts – through books, for example — that we didn’t know existed, just might be the key to helping us all think in new ways, see the world through a different lens, and see new ways to solve old and sticky problems.

So my hope for this weekend is that you hack with an eye to using the tools, the data, and talent assembled here to promote more accessible and equitable forms of serendipity.  I think our democracy, our world, could really use it right now.

 

 

 

 

Early reading list on machine learning

In the preliminary report from the MIT Task Force on the Future of Libraries, we make several references to the importance of optimizing library content, data, and metadata for machine learning applications.

We imagine a repository of knowledge and data that can be exploited and analyzed by humans, machines, and algorithms. This transformation will accelerate the accumulation and validation of knowledge, and will enable the creation of new knowledge and of solutions to the world’s great challenges. Libraries will no longer be geared primarily to direct readers but instead to content contributors, community curators, text-mining programs, machine-learning algorithms, and visualization tools.

I am convinced that machine learning is going to have a major impact on the advancement of knowledge in lots of ways we can’t anticipate, and I want to understand it better. I am also convinced that without the intervention of folks who understand the biases built into our collections in terms of content, organization, and description; machine learning applications will re-inscribe and reify existing inequalities.

To that end, I’m trying to put together a reading list to get smarter about what machine learning is, what it can do for libraries, and what libraries can do to support and inspire creative, productive, just and inclusive applications of machine learning. Here’s my very incomplete initial list. Additional suggestions welcome in the comments.

• Machine Learning: The new AI  by Ethem Alpaydin (part of the Essential Knowledge Series from MIT Press)
• Always Already Computational: Library Collections as Data, an IMLS-funded project, Thomas Padilla.
• “New AI-Based Search Engines are a “Game Changer” for Science Research”, by Nicola Jones, Nature magazine on November 12, 2016
• Artificial Intelligence Is Lost in the Woods, by David Gelernter, in MIT Technology Review
• Searching for Lost Knowledge in the Age of Intelligent Machines, by Adrienne Lafrance, The Atlantic.
• AI Songsmith Cranks Out Surprisingly Catchy Tunes, by Will Smith, MIT Technology Review
• And this list of resources: Hitchhiker’s guide to data science, machine learning, R, Python, from Vincent Granville

Libraries, technology, and social justice

Here’s the text of the talk I gave at Access 2016. I reused some stuff from earlier talks, but there’s some new stuff in here too. There is a video of the talk too.

(argh. I spelled Bethany Nowviskie’s name wrong on the slide in my talk. I hope she doesn’t notice.)

~~~~~~~~~~~~~~~~~~~~~~

Thank you for inviting me to this beautiful location and to this fantastic gathering. I want to give a special shout-out to James Mackenzie and the program committee for inviting me and for taking care of all the logistics of getting me here and especially for answering all my questions.

When I am asked to speak at conferences, I try to remember to ask a set of questions that include:

Do you have a code of conduct?

Do you have scholarships for people who might not otherwise be able to attend?

Are you making efforts to ensure diversity in attendance and a diverse line-up of speakers, panelists, presenters?

Access was a YES on all 3.

Fredericton is a very beautiful place

And Fredericton earned bonus points on the secret private criteria I use, which is “is it in an interesting and beautiful location?”

So it was really a no-brainer and I am thrilled to be here; and to have a chance to talk to and with all of you.

I want to start by saying that I’m so glad that Dr. Maclean acknowledged that the land on which we gather is traditional unceded territory.

The importance of acknowledging that we work on lands that are the traditional territories of First Nations people is something I am learning from my Canadian colleagues and from my Native American colleagues. It is, I think, a much needed way of showing recognition of and respect for aboriginal peoples.

I will say though, that it is a practice that is not as widespread in the US – yet.

But there is some movement in the US among colleges and universities to wrestle with their racist pasts; to acknowledge the role of slavery, and the mistreatment of native americans in their founding and early success.

Dozens of American universities – including Harvard, Brown, Columbia, Georgetown, and UVa — are in the process of publicly acknowledging their ties to slavery, including their dependence on slave labor; and these schools are beginning the work of trying to find paths to restitution, if not full reparations.

And this is not unrelated to the topic of my talk.

If I recall correctly, the abstract of this talk proclaims that libraries aren’t neutral, that technology isn’t neutral; and that we can and should leverage both in the service of social justice.

I figured I should spend at least a bit of time unpacking the claims that neither libraries nor technologies are neutral.

And one place to start for libraries – for academic libraries – is to acknowledge that our parent institutions are not and never have been neutral.

My point of reference is US colleges and universities, but I suspect the general theme is true in a Canadian context as well.

American colleges were originally built as exclusive institutions for well-connected white men; and in many cases American universities were actually built literally on the backs of enslaved African-american labor. Many of our institutions were built on land taken from native peoples; and almost all of our colleges and universities excluded in practice if not also in policy, women, non-white men, queer people, and other marginalized populations.

We start from these histories of exploitation, appropriation, enslavement, and displacement. And I believe we have a responsibility to acknowledge that we give our labor to institutions with often troubled histories with regard to the treatment and acceptance of women and non-white men. And even acknowledging that is a political act – but/and ignoring that past is also a political act. There is no neutral here.

In the US context I think it is important to give credit to the students – predominantly students of color — who came together on campuses across the country last fall, and continue to come together, to demand that universities own up to the systemic racism that exists in higher education and across the US, and to insist that we take steps to reduce discrimination and promote social justice.

Many of you likely heard about the high visibility student protests at the University of Missouri, and at Yale University; but in reality students and community members at hundreds of colleges across the nation took up the call and protested and demanded action from their own schools.

At MIT and at colleges all across the country, students have called our attention to ubiquitous and blatant incidents of racial and sexual harassment, they have demanded that we hire more faculty from underrepresented groups, and they have called for faculty and staff to be educated on unconscious bias.

In short, they have said – it isn’t enough to welcome students from marginalized groups to our campuses with words and policies; we must take concrete action to create welcoming, inclusive, and integrated communities. And in some cases, they have called on us to leverage the academy and its resources to address society level failings.

So what does that mean for us?

Well, that’s exactly what I want to talk about today – as folks who work in and around library technologies, how can we leverage our work in the service of social justice?

First, what is social justice and what does it look like?

I’m going to cheat a bit with the answer to what does social justice look like and cite a couple of things I’ve written or co-written in the past:

In an article titled Diversity, Social Justice & the future of Libraries that I had the honor of writing with Myrna Morales and Em Claire Knowles, we defined social justice as:

“The ability of all people to fully benefit from economic and social progress and to participate equally in democratic societies.”

If you believe like I do, that equitable access to information and to the tools to discover, use and understand that information; is a core enabling feature of a truly democratic society; then it is easy to see that libraries are crucial to social justice.

What would a social justice agenda look like in a library?

I was asked several months ago in a joint keynote I gave with my colleague Lareese Hall, now dean of libraries at the very prestigious Rhode Island School of Design, what a queer feminist agenda for libraries would look like, and I think that answer stands for a general social justice agenda too:

“A … feminist and queer agenda in an academic library would be one where the collections and services are not centered on the experiences of cis-straight, white western men; where the people who work in the library truly do reflect the diversity of the communities they serve; where the staff and patrons are empowered; and where the tools, systems, and policies are transparent and inclusive.”

For this crowd, at this conference; I want to talk about tools and technologies.

First, let me run through a few examples to illustrate what I mean when I say technology is not neutral; and really to convince any skeptics that technology itself – not just the users of it – is often biased.

Let’s start with search technologies. Most librarians will agree that commercial search engines are not “neutral” in the sense that commercial interests and promoted content can and do impact relevancy

But of course, declaring one thing as more relevant than another is always based on some subjective judgement – even if that judgment is coded into an algorithm many steps away from the output.

Or, as my colleague Bess Sadler says, the idea of neutral relevance is an oxymoron (this is a line from our Feminism and the future of library discovery article).

And of course, you can’t talk about bias in search tools without talking about the fantastic work of another one of my library sheros: Safiya Noble.

Safiya Noble’s work demonstrates how the non-neutrality of commercial search engines reinforce and perpetuate stereotypes, despite the fact that many of us assume the “algorithm” is neutral.

What Noble’s analysis of Google shows us is that Google’s algorithm reinforces the sexualization of women, especially black and Latina women. Because of Google’s “neutral” reliance on popularity, page rank, and promoted content, the results for searches for information on black girls or Latina girls are dominated by links to pornography and other sexualized content.

Noble suggests that users “Try Google searches on every variation you can think of for women’s and girls’ identities and you will see many of the ways in which commercial interests have subverted a diverse (or realistic) range of representations.”

So rather than show you those results, I encourage those of you who might be skeptical to do some of those searches yourself – google Asian girls, or latina girls, or black girls or native girls. And then Imagine being a girl or woman of color looking for yourself and your community on the web.

Or, just imagine you’re a tech worker

We know that the stereotype of a “tech worker” is young, male, nerdy … and the google image search verifies and reinforces that.

Google image search for ‘Tech worker” is pretty much all dudes

And labels matter – look at the different images you get when you search for “Librarian” vs. “Information scientist”

Google image search for librarian

Google image search for information scientist

 

We all like to think that library search tools can do better – and they can; but only when we are intentional about it.

Another example of technology that isn’t neutral comes from cameras and photo editing software.

Photographer Syreeta McFadden has written about how color film and other photographic technologies were developed around trying to measure the image against white skin.

The default settings for everything from film stock to lighting to shutter speed were and are designed to best capture “normal” faces – that is faces with white skin. What that means is that it is difficult to take photos of non-white faces that will be accurately rendered without performing post-image adjustments that sacrifice the sharpness and glossy polish that is readily apparent in photos of white faces.

And how many of you heard about the Twitter Bot that Microsoft created that became a crazy sexist racist idiot in less than 24 hours?

Last Spring, Microsoft unveiled a twitter bot named Tay; programmed to tweet like a teen. What could go wrong, right?

Tay is backed by Artificial Intelligence algorithms that were supposed to help the bot learn how to converse naturally on twitter. But what happened is that the bot learned quickly from the worst racist sexist corners of twitter – and within 24 hours Microsoft had to shut the experiment down because the bot had started tweeting all kinds of sexist, racist, homophobic, anti-Semitic garbage. Again, use your google skills to find them, I’m not sharing them from the podium.

For me the Microsoft experiment with a machine-learning twitterbot is a stark example of the fact that passive, mythical neutrality is anything but neutral. And sure you can blame it on the racist creeps on twitter, but creating technology that fails to anticipate the racist and sexist ways that technology might be used and exploited is not a neutral act. And I would venture to guess that it was a choice made by people who are least likely to have been the targets of discriminatory crap on the internet.

My bigger point here is that while crowd-sourcing and leveraging the social web are hot trends now in tech, I want to encourage us to think hard and critically about the consequences. Basically, I think we need to be very aware of the fact that if we crowd-source something, or if we rely on the social web or the sharing economy; we have to at least try to correct for the fact that the crowd is racist and sexist, and homophobic, and discriminatory in a whole bunch of horrifying ways.

There are all these great new services, that are part of what we call the Sharing economy that eliminate the “middle-man” and let people sell services directly to other people – to share things like rides and rooms with strangers. So there are ride-sharing apps like Uber and Lyft and services like Airbnb, where you can avoid hotels and hotel prices and stay in someone’s spare bedroom.

Stories abound in the US of Uber & Lyft drivers refusing to pick up passengers in minority neighborhoods, or canceling rides when they learn that a passenger is disabled and requires accommodations or assistance.

But I find the case of Airbnb especially interesting, because they are trying to fix their racism problem with both policy and technology.

So here’s what happened with AirBnB – first there was an experimental study out of Harvard about a year ago showing that renters were less likely to rent to people with black sounding names; then there were several reports of renters cancelling bookings for black guests; only to then rent to white guests for the same time period.

Honestly, this shouldn’t surprise us – the amount of social science evidence confirming that people act in biased ways in a huge variety of settings is overwhelming. What is interesting is that AirBnB is trying to do something about it, and they are being unusually transparent about it; so we might learn what works and what doesn’t.

First, they are having everyone who participates as a renter or a host sign a community agreement to treat everyone with respect and without bias. And there is some evidence that community compacts introduce some mutual accountability that has some positive effects, so that’s a good start. They are also providing training on unconscious bias to hosts and highlighting the hosts who complete the training on their website – which is a decidedly not neutral way of driving more renters to hosts who have completed the training.

What’s really interesting is that they are also working on technical features to try to eliminate instances where hosts claim a room or house is booked when a black renter makes a request; only to then immediately rent for the same time period to a white renter. Here is how they explain it: With the new feature If a host rejects a guest by stating that their space is not available, Airbnb will automatically block the calendar for subsequent reservation requests for that same trip.

They are also adding new flagging tools so people can report discrimination and hate speech.

And they have a team of engineers, data scientists, and designers who are looking for other ways to mitigate discrimination and bake some anti-bias features into their platform.

Would it have been better if they had anticipated the racist behavior enabled by their platform? Sure. But now that they are trying to make corrections, and to use technology to do it, I think there might be a real opportunity for us all to learn how we might leverage technology in combatting discrimination.

So, I’ve given some examples of how technology itself is not neutral. My point with these examples is to convince you that technology does not exist as neutral artifacts and tools that might sometimes get used in oppressive and exclusionary ways. Rather, technology itself has baked-in biases that perpetuate existing inequalities and exclusions, and that reinforce stereotypes.

How do we do not just try to mitigate the bias but also actually bring a social justice mindset to our work in library technology?

How do we promote an inclusive perspective, and an agenda of equity in and through our tech work?

First, we do everything we can to make sure the teams we have working on our tools and technologies and projects are actually inclusive and diverse.

And that is admittedly hard; but we do know some things that work. And by know, I mean there are actual scholarly studies that produce some evidence of practices that for example, discourage women from pursuing tech careers or applying for jobs. If I told you of a couple of simple things you could do that have shown they would remove some social barriers to women pursuing tech careers, would you be willing to do them?

(I stopped and waited until most of the room nodded their heads yes)

OK – here goes.

First things first – Don’t be this guy.

Don’t be the guy who says: “Always code as if the guy who ends up maintaining your code will be a violent psychopath who knows where you live.”

Don’t share advice like this; and don’t talk like this or joke like this.

This is some of the most horrendous advice about anything I have ever seen – or at least the worst I’ve seen about coding. And quite frankly I am certain it was written by someone who has a blind spot about the fact that women have to worry about being doxed by violent psychopaths just for being on the internet; or being stalked, attacked and too often killed for ignoring the advances of strangers, or for confronting cat-callers. Queer and trans people are also overwhelmingly more likely to be victims of violent crimes; especially trans women of color.

So using, even in jest, the specter of a violent psychopath, to encourage good coding practices is not just a crappy thing to do – it also reinforces a culture that is hostile to women and to other marginalized groups.

And I know we don’t want to admit it, but technology has a culture problem – even in libraries. Remember those search results for “tech worker” – they reflect the predominant image of who works in technology.

So what are some ways we can make technology work more inclusive?

I want to talk about 3 ways:

1. change the image of the “tech guy”
2. change the work environment
3. watch your language (but not in the way you might think)

First, let’s talk about the “tech guy” image.

Some colleagues of mine at Stanford, sociologists Alison Wynn and Shelley Correll, have done some very interesting work looking at how well people who are already working & succeeding in technology jobs felt they matched the cultural traits & stereotypes of a successful tech worker; and how that sense of a match, or in the case of most women, the sense of a mismatch, effects a number of outcomes. (I don’t have a citation for this study, because it is still under review for publication. Because Shelley is an old friend, I knew about the research and got to read an unpublished version; which she gave me permission to reference in talks, but no citation. Scholarly communication is broken.)

First they developed a composite scale based on how tech workers, men and women, described successful tech workers. Ask people to come up with some adjectives to describe a “successful tech worker” and not too surprisingly the stereotype that emerged was masculine, obsessive, assertive, cool, geeky, young, and working long hours. In other words, The “Tech guy” stereotype is wide-spread and well-known.

And as we would expect, their data show that women tech workers are significantly less likely than their male counterparts to view themselves as fitting that cultural image of a successful tech worker.  Where it gets interesting though is that their research goes on to show that the sense of not fitting the cultural image has consequences.

Because women are less likely to feel they fit the image of a successful tech worker, they are less likely to identify with the tech field, more likely to consider leaving the tech field for another career, and less likely to report positive treatment from their supervisors.

Reminder that their sample was men and women currently working in tech jobs in silicon valley tech firms. So successful women in tech see themselves as not fitting in; and as a result are leaving the field.

The bottom line is that cultural fit matters – not just in the pipeline, as women decide whether to major in STEM fields or to pursue tech jobs – but also among women who are currently working in technology. In other words, stereotypes about tech work and tech workers continue to hinder women even after they have entered tech careers. If we want to ensure that our technologies are built by diverse and inclusive groups of people, we have to find ways to break down the stereotypes and cultural images associated with tech work.

And that brings us to the Star Trek posters – which is somehow always the most controversial part of talks I give on this topic.

But let’s get to the research — In a fascinating experimental study, psychologist Sapna Cheryan and colleagues found that women who enter a computer science environment that is decorated with objects stereotypically associated with the field – such as Star Trek posters or video games– are less likely to consider pursuing computer science than women who enter a computer science environment with non-stereotypical objects — such as nature or travel posters. These results held even when the proportion of women in the environment was equal across the two differently decorated settings.

The Star Trek posters and other seemingly neutral nerdy dude paraphernalia we use to decorate our communal tech spaces serve to deter women – and I expect some of it deters men from marginalized groups as well.

So, to sum up – we can make tech more inclusive if we stop using the term “tech guy”, if we try to promote images of tech workers that aren’t just geeky, obsessive dudes who work long hours, and if we get rid of the Star Trek posters in our communal & public spaces.

And I know some of you are thinking “but I like my Star Trek posters”, but I hope your commitment to diversity wins out over your devotion to your Star Trek posters. Because increasing the number of women in tech is hard, and we have very little research to guide us; but we do know that the Star Trek stuff makes tech work less appealing to women.

And finally, watch your language.

Research also shows that certain words in job ads discourage women from applying. Research shows that women are less likely to apply for engineering and programming jobs when those ads have stereotypically masculine words like “competitive” or “dominate”. Women are less likely to apply and are more likely to feel that they wouldn’t fit in or belong when words like that are part of the job description. This is a case where technology can help – there are text analysis programs that can tell you if you are using gendered language in your job ads and can suggest more neutral language.

But again, this just points to the fact that if we want our technology to work towards diversity, inclusion and equity; we have to intervene and design it explicitly to do so.

That’s one of the lessons learned by a set of researchers who trained a machine learning algorithm on Google news articles then asked the algorithm to complete the analogy:

“Man is to Computer Programmer as Woman is to X.” The answer came back: “Homemaker.”

In fact, when asked to generate a large numbers of

He is to X as She is to Y analogies, the algorithm returned plenty more stereotypes:

• He is to doctor as She is to nurse
• He is to brilliant as She is to lovely
• He is to pharmaceuticals as She is to cosmetics

The corpus of text the machine learning algorithm learned on was itself biased and filled with stereotypes and stereotypical associations.

But again, there are ways to de-bias the system using human intervention.

In this case, a team of researchers flagged associations the algorithm had made that were gendered and added code instructing the algorithm to remove those associations. The algorithm could be taught to recognize and remove bias.

OK – I started off with the notion that libraries aren’t neutral and technology is not neutral; and I’ve talked about lots of examples of technologies that aren’t neutral either in their design or in their execution or both. And I’ve offered some research to help bring more diversity to our library technology teams, in the hope that more diverse and inclusive teams building our technologies will lead to design choices that favor social equity and justice.

But let me be clear – I don’t think increasing the percentage of women, and men of color in our technology departments is a magic bullet and I certainly don’t think we need to wait until we are more diverse to start thinking about how to leverage our technology work to promote social justice. I think we need to increase the diversity of our libraries, in technology and throughout the profession – but numbers aren’t the only answer.

I have some general ideas about how we might build library technologies for social justice and I’ll share them quickly because I want to hear your ideas.

First, I think we need to consciously think about social justice principles and try to build them into every step of our work. For me social justice principles are feminist principles – transparency, participation, agency, embodiment. We should also ask who is missing from our work, or from the personas we develop. And if the answer is women; then we need to dig deeper and ask which women? Too often we think adding white women fixes our diversity problem.

If we really want to work on tech projects that promote social justice in our communities then we need to talk to our most marginalized community members. At my institution, that would be the racial and ethnic identity student groups, the queer and the trans students, the Muslim students. If we reach out to these groups specifically and try to find out what they need, what they struggle with in the library and more generally at our institutions, we might realize that there are technology projects that would help.

And in all of our work, I think we get closer to social justice the more we practice the art of truly listening to each other and to our communities.

I also want to promote an ethic of care and empathy which is something 2 of my favorite humanists have recently written about: Bethany Nowviskie, executive director of DLF wrote about this in a piece titled “on capacity and care”; and just this weekend Kathleen Fitzpatrick, president of the Modern Language Association wrote about a new project she is calling “Generous thinking.” I recommend both to you.

And in that spirit of listening, it is time for me to wrap this up and to hear from you. I hope you will feel free to say whatever you want, to make comments of all kinds, no need to phrase it in the form of a question. A conversation among all us is much more interesting than me answering questions. So I’m ready to listen now. Thank you

Never neutral: Libraries, technology, and inclusion

Below is the text from the OLITA Spotlight talk I gave at the OLA Super Conference (#olasc15).

~~~~~~~~~~~~~

I want to acknowledge from the outset that this talk has been heavily influenced by a number of people who have shared their work and their thoughts with me over the years. I’ve been privileged to learn from them, in some cases formally through their publications and in some cases through conversations on twitter or even in person. These aren’t the only folks whose work and thinking influences me, but they are the key people I think of when I think of critical work on the intersections of libraries, technology, higher education and social justice.  These are their names – a mix of students, librarians, scholars, and technologists. Again, this is not a comprehensive list of the people whose work inspires me, but they are my top 7 right now on these topics.

• nina de jesus
• Myrna Morales
• Cecily Walker
• Bess Sadler
• Tressie McMillan Cottom
• Safiya Noble
• Andromeda Yelton

Let me also acknowledge that I’m well aware that the fact that I am a white woman working at an elite private US university gives me access to a platform like this one to talk about issues of bias and exclusion in libraries and technology. But there are plenty of folks who have been and continue to talk about and write about these issues, with far more insight and eloquence than I can, but who don’t get invitations like this for a variety of reasons. And the sad truth is that what I say, as an associate director at Stanford Libraries or as Director of MIT Libraries, often gets more attention than it deserves because of my title; while folks with less impressive titles and less privilege have been talking & thinking about some of these issues for longer than me and have insights that we all need to hear.

So next time you are looking for a speaker, please consider one of the names listed above.

If you read the blurb describing this talk, you know that a fundamental tenet that undergirds this talk, and frankly undergirds much of the work I have done in and for libraries, is the simple assertion that libraries are not now nor have they ever been merely neutral repositories of information. In fact, I’m personally not sure “neutral” is really possible in any of our social institutions … I think of neutral as really nothing more than a gear in your car.

Title slide for Never Neutral talk

But what I mean when I say libraries are not neutral is not just that that libraries absorb and reflect the inequalities, biases, ethnocentrism, and power imbalances that exist throughout our host societies and (for those of us who work in academic libraries) within higher education.

I mean that libraries are not neutral in a more direct and active way.

For an exceptionally compelling take on libraries as not just not neutral, but as instruments themselves of institutional oppression, please read “Locating the Library in Institutional Oppression” by my friend and colleague nina de jesus.

nina argues that “Libraries as institutions were created not only for a specific ideological purpose, but for an ideology that is fundamentally oppressive in nature.” It is a bold argument, convincingly made; and I urge you to read it. As a bonus, the article itself is Open Access and nina elected to use only Open Access sources in writing it.

So I start with the premise that it isn’t just that libraries aren’t perfectly equitable or neutral because we live in a society that still suffers from racism, sexism. ableism, transphobia and other forms of bias and inequity; but libraries also fail to achieve any mythical state of neutrality because we contribute to bias and inequality in scholarship, and publishing, and information access.

Let me step back for a minute and own up to a few of my own biases – my library career thus far has been solely and squarely within large academic libraries; so my perspective, my examples, and my concerns come out of that experience and are likely most relevant to that sector of libraries. But, I hope we can have a conversation at the end of my talk about what the differences and similarities might be between the way these issues play out in large academic libraries and the way they play out in all kinds and sizes of libraries. I’m also definitely speaking from an American perspective, and I look forward to hearing where and how cultural differences intersect with the ideas I’ll talk about.

OK – so libraries are not neutral because we exist within societies and systems that are not neutral. But above and beyond that, libraries also contribute to certain kinds of inequalities because of the way in which we exercise influence over the diversity (or lack thereof) of information we make available to our communities and the methods by which we provide access to that information.

I have a whole other talk that I’ve given on how the collection development decisions we make impact not just how inclusive or not our own collections are, but also what kinds of books and authors and topics get published. The short version of that talk is that when we base our purchasing decisions on circulation and popularity, we eliminate a big part of the market for niche topics and underrepresented authors. That is bad for libraries, bad for publishing, and bad for society. But that’s another talk. This talk is about library technologies.

But before we get into technology per se., I think a word about our classification systems is necessary, because the choices we make about how our technologies handle metadata and catalog records have consequences for how existing biases and exclusions get perpetuated from our traditional library systems into our new digital libraries.

Many of you are likely well aware of the biases present in library classification systems.
Hope Olson – one of the heroes of feminist and critical thinking in library science – has done considerable work on applying critical feminist approaches to knowledge organization to demonstrate the ways in which libraries exert control over how books and other scholarly items are organized and therefore how, when, and by whom they are discoverable.

Our classification schemes — whether Dewey Decimal  or Library of Congress — are hierarchical, which leads to the marginalization of certain kinds of knowledge and certain topics by creating separate sub-classifications for topics such as “women and computers” or “black literature”.

Let me give a couple of examples of the effects of this.

Call numbers matter

The power of library classification systems is such that a scholar browsing the shelves for books on military history is unlikely to encounter Randy Shilts’ seminal work Conduct Unbecoming: Gays & Lesbians in the US Military, because that book has been given a call number corresponding to “Minorities, women, etc. in armed forces”.  In my own library at Stanford University, that means the definitive work on the history of gays and lesbians serving in the armed forces is literally shelved between Secrets of a Gay Marine Porn Star and Military Trade — a collection of stories by people with a passion for military men.  Now I’m not saying we shouldn’t have books about gay military porn stars or about those who love men in uniform. I am saying that there is nothing neutral about the fact that the history of gay & lesbian service members is categorized alongside these titles, while the history of “ordinary soldiers” (that’s from an actual book title) is shelved under “United States, History – Military.”

Another example is one I learned of from my friend and colleague Myrna Morales, and you can read about it in an article I co-authored with her and Em Claire Knowles. In that article, Myrna writes about her experience doing research for her undergraduate thesis on the Puerto Rican political activism that took place in NYC in the 1960s, with a special interest in the Young Lords Party.

Here is how Myrna described her experience:

I first searched for the YLP with the subject heading “organizations,” subheading “political organization,” in the Reader’s Guide to Periodical Literature. Here I found no mention of the YLP. I was surprised, as I had known the YLP to be a prominent political organization—one that addressed political disenfranchisement, government neglect, and poverty. A (twisted) gut feeling told me to look under the subject heading of “gangs.” There it was—Young Lords Party. This experience changed my view of the library system, from one impervious to subjectivity and oppression to one that hid within the rhetoric of neutrality while continuing to uphold systemic injustices.

I suspect that this kind of experience is all too common for people of color and other marginalized people who attempt to use the resources we provide. I’ll go so far as to wonder if these sorts of experiences aren’t at least partially responsible for the incredibly low proportion of people of color who pursue careers in librarianship.

So our traditional practices and technologies are not neutral, and without active intervention we end up with collections that lack diversity and we end up classifying and arranging our content in ways that further marginalizes works by and about people of color, queer people, indigenous peoples, and others who don’t fit neatly into a classification system that sets the default as the as western, white, straight, and male.

Of course, the promise of technology is that we no longer need rely on arcane cataloging rules and browsing real library stacks to discover and access relevant information. With the advent of online catalogs and search engines, books and other information items can occupy multiple “places” in a library or collection.

But despite the democratizing promise of technology, our digital libraries are no more capable of neutrality than our traditional libraries; and the digital tools we build and provide are likely to reflect and perpetuate stereotypes, biases, and inequalities unless we engage in conscious acts of resistance.

Now when most people talk about bias in tech generally or in library technology, we talk about either the dismal demographics that show that white women and people of color are way underrepresented in technology, or we talk about the generally misogynistic and racist and homophobic culture of technology; or we talk about both demographics and culture and how they are mutually reinforcing. What we talk about less often is this notion that the technology itself is biased – often gendered and/or racist, frequently ableist, and almost always developed with built in assumptions about binary gender categories.

For some folks, the idea that technologies themselves can be gendered, or can reflect racially based and/or other forms of bias is pretty abstract. So let me give a few examples.

Most librarians will agree that commercial search engines are not “neutral” in the sense that commercial interests and promoted content can and do impact relevancy. Or, as my colleague Bess Sadler says, the idea of neutral relevance is an oxymoron.

Safiya Noble’s work demonstrates how the non-neutrality of commercial search engines reinforce and perpetuate stereotypes, despite the fact that many assume the “algorithm” is neutral.

What Noble’s analysis of Google shows us is that Google’s algorithm reinforces the sexualization of women, especially black and Latina women. Because of Google’s “neutral” reliance on popularity, page rank, and promoted content, the results for searches for information on black girls or Latina girls are dominated by links to pornography and other sexualized content. Noble suggests that users “Try Google searches on every variation you can think of for women’s and girls’ identities and you will see many of the ways in which commercial interests have subverted a diverse (or realistic) range of representations.”

Search technologies are not neutral – just as basing collection development decisions on popularity ensures that our collections reflect existing biases and inequalities, so too does basing relevancy ranking within our search products on popularity ensure the same biases persist in an online environment.

But it isn’t just search engines. In an article called “Teaching the Camera to see my skin”, photographer Syreeta McFadden describes how color film and other photographic technologies were developed around trying to measure the image against white skin. Because the default settings for everything from film stock to lighting to shutter speed were and are designed to best capture white faces; it is difficult to take photos of non-white faces that will be accurately rendered without performing post-image adjustments that sacrifice the sharpness and glossy polish that is readily apparent in photos of white faces.

Teaching the camera to see my skin

Finally, in an example of a technology that betrays its lack of neutrality by what it ignores, Apple’s recently released health app allows users to track a seemingly endless array of health and fitness related information on their iPhone. But strangely, Apple’s health app did not include a feature for tracking menstrual cycles – an important piece of health data for a huge percentage of the population. As one critic noted, Apple insists that all iPhone uses have an app to track Stock prices – you can’t delete that one from your phone — but fails to provide an option for tracking menstrual cycles in its “comprehensive” health tracking application.

I hope these examples demonstrate that technology does not exist as neutral artifacts and tools that might sometimes get used in oppressive and exclusionary ways. Rather, technology itself has baked-in biases that perpetuate existing inequalities and exclusions, and that reinforce stereotypes.

So how do we intervene, how do we engage in acts of resistance to create more inclusive, less biased technologies?

Note that I don’t think we can make completely neutral technologies … but I do think we can do better.

One way we might do better is simply by being aware and by asking the questions that the great black feminist thinkers taught us to ask:

Who is missing?

Whose experience is being centered?

Many, many folks argued – rather convincingly to my mind – that the dearth of women working at Apple may have contributed to the company’s ability to overlook the need for menstrual cycle tracking in its health app.

So we might also work on recruiting and retaining more white women and people of color into library technology teams and jobs. There is much good work being done on trying to increase the diversity of the pipeline of people coming into technology – Black Girls Code and the Ada Initiative are examples of excellent work of this type.

I also think the adoption of strong codes of conduct at conferences like this one and other library and technology events make professional development opportunities more welcoming and potentially safer for all – and I think those are important steps in the right direction.

But in the end, one of the biggest issues we need to address if we truly want a more diverse set of people developing the technologies we use is the existence of a prevailing stereotype about who the typical tech worker is.

I want to turn now to some research on how stereotypes about who does technology, and who is good at it, affect how interested different kinds of people are in pursuing technology related fields of study, how well people expect they will perform at tech tasks, and how well people already working in tech feel they fit in, and how likely they are to stay in tech fields.

First a definition – Stereotypes are widely shared cultural beliefs about categories of people and social roles. The insidious thing about stereotypes is that even if we personally don’t subscribe to a particular stereotype, just knowing that a stereotype exists can affect our behavior.

Second, a caution – much of this research focuses on gender, to the exclusion of intersecting social identities such as race, sexuality, or gender identity. The research that talks about “women’s” behavior and attitudes towards technology is usually based on straight white women .. so keep that in mind, and recognize that much more research is needed to capture the full range of experiences that marginalized people have with and in technology.

That said, there is a huge body of research documenting the effect of negative stereotypes about women’s math and science abilities. These kinds of stereotypes lead to discriminatory decision making that obstructs women’s entry into and advancement in science and technology jobs. Moreover, negative stereotypes about women and math affects women’s own self-assessment of their skill level, interest, and suitability for science and technology jobs.

Barbie “Math is hard”

In a not yet published research study of men and women working in Silicon Valley technology firms, Stanford sociologists Alison Wynn and Shelley Correll looked at the impact of how well tech workers felt they matched the cultural traits of a successful tech worker on a number of outcomes.

First they developed a composite scale based on how tech employees, men and women, described successful tech workers. The stereotype that emerged was masculine, obsessive, assertive, cool, geeky, young, and working long hours.

Their data show that women tech workers are significantly less likely than their male counterparts to view themselves as fitting the cultural image of a successful tech worker.  While that may not be a surprising finding, their research goes on to show that the sense of not fitting the cultural image has consequences.

Because women are less likely to feel they fit the image of a successful tech worker, they are less likely to identify with the tech field, more likely to consider leaving the tech field for another career, and less likely to report positive treatment from their supervisors.

The bottom line is that cultural fit matters – not just in the pipeline, as women decide whether to major in STEM fields or to pursue tech jobs – but also among women who are currently working in technology. In other words, stereotypes about tech work and tech workers continue to hinder women even after they have entered tech careers. If we want to ensure that our technologies are built by diverse and inclusive groups of people, we have to find ways to break down the stereotypes and cultural images associated with tech work.

How do we do that?

If we want to look to success stories, Carnegie Mellon University is a good example. At Carnegie Mellon they increased the percentage of women majoring in computer science from 7% in 1995 to 42% in 2000 by explicitly trying to change the cultural image of computer scientists. Faculty were encouraged to discuss multiple ways to be a computer scientist and to emphasize the real world applications of computer science and how computer science connects to other disciplines. They also offered computer science classes that explicitly stated that no prerequisites in math or computer science were required.

For libraries, we can talk about multiple ways to be a library technologist, and we can emphasize the value of a wide variety of skills in working on library tech projects – metadata skills, user experience skills, design skills. We can provide staff with opportunities to gain tech skills in low-threat environments and in environments where white women and people of color are less likely to feel culturally alienated.

RailsBridge workshops and AdaCamps seem like good fits here, and I’d like to see more library administrators encouraging staff from across their org’s to attend such training. At Stanford, my colleagues Bess Sadler and Cathy Aster started basic tech training workshops for women on the digital libraries’ staff who were doing tech work like scanning, but who didn’t see themselves as tech workers. Providing the opportunity to learn and ask questions, in a safe environment away from their supervisors and male co-workers gave these women skills and confidence that enhanced their work and the work of their groups.

Another simple way we can make progress within our own organizations is to pay attention to the physical markers of culture.

In a fascinating experimental study, psychologist Sapna Cheryan and colleagues found that women who enter a computer science environment that is decorated with objects stereotypically associated with the field – such as Star Trek posters — are less likely to consider pursuing computer science than women who enter a computer science environment with non-stereotypical objects — such as nature or travel posters. These results held even when the proportion of women in the environment was equal across the two differently decorated settings.

We need to pay attention to the computer labs and maker spaces in our libraries, and we need to pay attention to physical work environments our technical staff work in. By simply ensuring that these environments aren’t plastered with images and objects associated with the stereotypes about “tech guys”, we will remove one of the impediments to women’s sense of cultural fit.

So let me try to sum up here.

I’ve argued that like libraries, technology is never neutral. I’ve offered examples from search engines to photography to Apple’s health tracking app.

I’ve talked about how the pervasive stereotypes about who does tech work limit women’s participation in tech fields, through both supply and demand side mechanisms.

The stereotypes about tech workers also contain assumptions about race and sexuality in the US context, in that the stereotypical tech guy is white (or Asian) and straight. Sadly, there is significantly less research on the effect of those stereotypes on black and Latino men and women and queer people who are also vastly underrepresented in technology work.

Let me offer some parting thoughts on how we might make progress.

To borrow from the conference theme, we need to think and we need to do.

We need to think about the technology we use in our libraries, and ask where and how it falls short of being inclusive. Whose experiences and preferences are privileged in the user design? Whose experiences are marginalized? Then we need to do what we can to push for more inclusive technology experiences. We likewise need to be transparent with our patrons about how the technology works and where and how the biases built into that technology might affect their experience. The folks who do work in critical information literacy provide great models for this.

We should think about how libraries and library staff reinforces stereotypes about technology and technology work. Subtle changes can make a difference. We should drop the term “tech guy” from our vocabulary and we should ditch the Star Trek posters. I’d like to see more libraries provide training and multiple paths for staff to develop tech skills and to become involved in technology projects. We need to pay attention to the demographics and to the culture – and remember that they are mutually reinforcing.

We also need to remember that we aren’t striving for neutral, and we aren’t aiming for perfectly equitable and inclusive technology.

While neutral technologies are not possible – or necessarily desirable – I believe that an awareness of the ways in which technology embodies and perpetuates existing biases and inequalities will help us make changes that move us towards more inclusive and equitable technologies.

A thank you, a picture, and a call for witnesses

I had loads of fun helping to promote the #libs4ada fundraising drive, and am more than a bit amused to think that the promise of seeing me in a dress may have motivated a few extra dollars. I’m more than willing to make a fool of myself for a good cause.
If you missed the tweet with the actual #ButchInADress photo, here it is:

Butch in a dress

Huge thanks to all those who donated and/or supported the fundraising drive in some way. It was a delight to see librarians blow past all the goals and keep on giving. We really do have an amazing core of generous folks in this profession who care deeply about making libraries, technology, and library technology more inclusive, more equitable, and more welcoming; and are willing to support the work of the Ada Initiative in those efforts.

Another vitally important way to support women in our profession is to believe them, to support them, to back them up when they are brave enough to speak out about harassment. Lisa Rabey and nina de jesus are being sued for speaking out about harassment in the library community and are calling for witnesses. I hope those who can bear witness will do so.

“Dressing” for the cause #libs4ada

Librarians (and our friends) are amazing. In the first day of the #libs4ada Ada Initiative fundraising drive, we blew our original goal of $5120 out of the water.

In all the excitement, I issued two “stretch” challenges:

1. If/when donations reach $8192, I will post a photo of myself in a dress selected for me by my very fashionable daughter.
2. If/when donations reach $15,000, twitter librarians can select a dress which I will wear for 10 minutes at the upcoming DLF Forum. Who knows? Maybe we can even “sell” selfies with me in said dress as an additional fundraiser.

So, if you think #ButchInADress might be fun/amusing/terrifying, give generously.
Y’all rock.

Young butch in a dress, Chris Age 4

This librarian supports the Ada Initiative

The Ada Initiative supports women in open technology and culture through activities such as producing codes of conduct and anti-harassment policies, advocating for gender diversity, teaching ally skills, and hosting conferences for women in open tech/culture. Most of what we create is freely available, reusable, and modifiable under Creative Commons licenses.

If that isn’t enough to explain why I support the Ada Initiative and why I think other librarians should too, let me tell you just one story about how the Ada Initiative has been important to me.

A little over a year ago, I decided that I was not going to speak at or support any conference that did not have a code of conduct. Then, in a fit of bravada, I decided to ask my boss, the University Librarian, to issue a statement encouraging ALL of our librarians to take the same stance and to work with the professional associations they were involved with to adopt codes of conduct. Making my argument was easy, because the Ada Initiative folks had already compiled all the data and documentation, and examples. The boss said yes, and I know of several major conferences that have since adopted codes at least partially in response to advocacy from Stanford librarians. I am convinced that these conferences are now a little more welcoming to folks who might otherwise have felt less included and less safe. I’m proud of whatever small role Stanford Libraries may have played in that — and the groundwork done by the Ada Initiative made that possible.

The truth is, I don’t really do much tech myself, but I’m a leader in an organization and a profession that does a lot of tech, and that employs many women. I very much want library technology to be diverse, inclusive, and as equitable as we can possibly make it; and The Ada Initiative gives us the tools to move in that direction.

So I support the Ada Initiative, and I hope you will too. Please also help spread the word via the hashtag #libs4ada.

Some research on gender, technology, stereotypes and culture

Leading up to the Leadership, Technology, and Gender Summit, my colleague-friend Jennifer Vinopal and I have been collecting sets of recommended readings to help frame the conversations. We tried to find readings that address most of the topics outlined in What are we talking about when we talk about Leadership, Technology and Gender, while also keeping the list to a reasonable length and making sure the readings were accessible to all. It is a great list.

Here are some additional recommended readings that did not make the list for LTG, but which I think are critical to understanding the scope of the challenges involved in tackling the problems of gender and technology. Most of them are paywall, which sucks.*

“Do Female and Male Role Models Who Embody STEM Stereotypes Hinder Women’s Anticipated Success in STEM?” Sapna Cheryan, John Oliver Siy, Marissa Vichayapai, Benjamin J. Drury and Saenam Kim Social Psychological and Personality Science 2011 2: 656 originally published online 15 April 2011 DOI: 10.1177/1948550611405218

Abstract
Women who have not yet entered science, technology, engineering, and mathematics (STEM) fields underestimate how well they
will perform in those fields (e.g., Correll, 2001; Meece, Parsons, Kaczala, & Goff, 1982). It is commonly assumed that female role models improve women’s beliefs that they can be successful in STEM. The current work tests this assumption. Two experiments varied role model gender and whether role models embody computer science stereotypes. Role model gender had no effect on success beliefs. However, women who interacted with nonstereotypical role models believed they would be more successful in computer science than those who interacted with stereotypical role models. Differences in women’s success beliefs were mediated by their perceived dissimilarity from stereotypical role models. When attempting to convey to women that they can be successful in STEM fields, role model gender may be less important than the extent to which role models embody current STEM stereotypes.

“STEMing the tide: Using ingroup experts to inoculate women’s self-concept in science, technology, engineering, and mathematics (STEM).” Stout, Jane G.; Dasgupta, Nilanjana; Hunsinger, Matthew; McManus, Melissa A.
Journal of Personality and Social Psychology, Vol 100(2), Feb 2011, 255-270. doi: 10.1037/a0021385

Abstract
Three studies tested a stereotype inoculation model, which proposed that contact with same-sex experts (advanced peers, professionals, professors) in academic environments involving science, technology, engineering, and mathematics (STEM) enhances women’s self-concept in STEM, attitudes toward STEM, and motivation to pursue STEM careers. Two cross-sectional controlled experiments and 1 longitudinal naturalistic study in a calculus class revealed that exposure to female STEM experts promoted positive implicit attitudes and stronger implicit identification with STEM (Studies 1–3), greater self-efficacy in STEM (Study 3), and more effort on STEM tests (Study 1). Studies 2 and 3 suggested that the benefit of seeing same-sex experts is driven by greater subjective identification and connected- ness with these individuals, which in turn predicts enhanced self-efficacy, domain identification, and commitment to pursue STEM careers. Importantly, women’s own self-concept benefited from contact with female experts even though negative stereotypes about their gender and STEM remained active.

“Math–Gender Stereotypes in Elementary School Children” Dario Cvencek, Andrew N. Meltzoff, Anthony G. Greenwald Article first published online: 9 MAR 2011 DOI: 10.1111/j.1467-8624.2010.01529.x

A total of 247 American children between 6 and 10 years of age (126 girls and 121 boys) completed Implicit Association Tests and explicit self-report measures assessing the association of (a) me with male (gender identity), (b) male with math (math–gender stereotype), and (c) me with math (math self-concept). Two findings emerged. First, as early as second grade, the children demonstrated the American cultural stereotype that math is for boys on both implicit and explicit measures. Second, elementary school boys identified with math more strongly than did girls on both implicit and self-report measures. The findings suggest that the math–gender stereotype is acquired early and influences emerging math self-concepts prior to ages at which there are actual differences in math achievement.

“Ambient Belonging: How Stereotypical Cues Impact Gender Participation in Computer Science” Sapna Cheryan,Paul G. Davies, Victoria C. Plaut and Claude M. Steele. Journal of Personality and Social Psychology Vol. 97, No. 6, 1045–1060 DOI: 10.1037/a0016239

People can make decisions to join a group based solely on exposure to that group’s physical environment. Four studies demonstrate that the gender difference in interest in computer science is influenced by exposure to environments associated with computer scientists. In Study 1, simply changing the objects in a computer science classroom from those considered stereotypical of computer science (e.g., Star Trek poster, video games) to objects not considered stereotypical of computer science (e.g., nature poster, phone books) was sufficient to boost female undergraduates’ interest in computer science to the level of their male peers. Further investigation revealed that the stereotypical broadcast a masculine stereotype that discouraged women’s sense of ambient belonging and subsequent interest in the environment (Studies 2, 3, and 4) but had no similar effect on men (Studies 3, 4). This masculine stereotype prevented women’s interest from developing even in environments entirely populated by other women (Study 2). Objects can thus come to broadcast stereotypes of a group, which in turn can deter people who do not identify with these stereotypes from joining that group.

Other great resources include:
Women in Computer Sciences: Closing the Gender Gap in Higher Education from Carnegie Mellon University

Starting in 1995, we have engaged in an interdisciplinary program of research and action in response to this situation. The research effort has been to understand male and female students’ engagement — attachment, persistence, and detachment — with computer science, with a special focus on the gender imbalance in the field. Students in the study have been interviewed once per semester about their family and schooling history, experiences with computing, feelings and attitudes about studying computer science. The goal of the action component has been to devise and effect changes in curriculum, pedagogy and culture that will encourage the broadest possible participation in the computing enterprise.

In part as a result of our efforts, the entering enrollment of women in the undergraduate Computer Science program at Carnegie Mellon has risen from 8% in 1995 to 42% in 2000

Reading list for MIT Open CourseWare course “Gender and Technology”:

Course Description
This course considers a wide range of issues related to the contemporary and historical use of technology, the development of new technologies, and the cultural representation of technology, including the role women have played in the development of technology and the effect of technological change on the roles of women and ideas of gender. It discusses the social implications of technology and its understanding and deployment in different cultural contexts. It investigates the relationships between technology and identity categories, such as gender, race, class, and sexuality, and examines how technology offers possibilities for new social relations and how to evaluate them.

* If you need access to any of these articles for your personal non-profit, educational use, contact a librarian near you ;-)