Hello hello everyone! How is everyone enjoying this (chilly) spring break?
If you remember, last month I wrote about three super insightful books that I thought everyone should try reading and in that collection was one book in particular called Weapons of Math Destruction by Cathy O’Neil. If you missed the post, here’s a recap:
“Once a Wall Street analyst, O’Neil became aware of the toxic manipulation of math by big corporations and powerful individuals and decided to leave her position at a huge hedge fund to expose the dangers of large-scale data modeling. A large portion of our future is determined by mathematical models, and while at its surface this can seem like a fair and unbiased approach it’s crucial to realize that one size does not fit all. Big data facilitates large-scale decisions and in time has gained a huge amount of agency over our lives, yet models notoriously lack depth and indirectly target the most vulnerable groups of people. O’Neil thoroughly explains how these models affect us directly by using very contemporary situations, for example college admissions, online ads, and finances; she aims to get a reaction from you in order to demand better use of big data. […] it is a reminder that while technology and models are improving, they have the potential to be very harmful to the least powerful in our society.”
Cathy O’Neil, who runs MathBabe.org, is a mathematician, author/journalist, musician, and musician. O’Neil attended UC Berkeley for her undergrad career before earning her PhD in math from Harvard, and was in academia for 8 years before going into finance. After becoming frustrated with the toxic misuse of data she left finance to join the Occupy Wall Street movement and has been an advocate for social justice in big data since. I was feeling bold and thought I’d e-mail O’Neil asking for an interview, and to my surprise she was actually down for the cause! We talked about being a woman in math, her career transitions, and what we can all do to bring attention to the problem of big data. I hope you enjoy everything she has to say as much as I did!
Why did you decide you wanted to pursue math as a field of study?
I’ve always loved math, and I also love music so I’ve always been tempted to become a musician, but the summer after my freshman year of high school I found out about a math summer camp; it was a hippie math camp and I was a hippie girl. It was a very “math is beautiful, let’s love math together” type thing. It was so much fun, and I loved it so much that by the end of the summer I was like, I’m gonna become a mathematician, so I was fourteen when I decided. For me it’s a type of art. It’s like music and it deserves to be loved and appreciated.
Do you ever regret not pursuing music?
I do pursue music! I’m in a band and we have gigs every month in the West Village. Actually my parents are the ones who talked me into it, they said “If you’re a mathematician you can still play music, but if you’re a musician it’s hard to do math.” So I thought about that and said, yeah, I can have both, and I do. It’s fun, I’m lucky.
What is your experience like being a woman in this field, both when you were in college and today?
When I was in college I got an award for undergrad women in math called the Alice Schafer prize, and I honestly was like, oh good it’s money! I was like whatever, I’m not a woman mathematician I’m just a mathematician, and I did appreciate the people that I met when I was receiving the award, but I didn’t see myself as a woman in math until I got to grad school. I saw it, there was real sexism. My mother was a mathematician as well and got her PhD in Harvard but in the 60’s. It was tough, she was in the applied math department with two other women and both of them married their advisers and never finished, so she was the only woman who did. Women were not supposed to have a career, can you imagine? So my mother had to deal with a lot of crap, but I thought I didn’t until I did. So I think my answer is we need a Me Too movement in math; it’s not just Hollywood. I actually wrote a piece for Bloomberg saying that where the data isn’t, that’s where the problems are. We don’t hear about women in other fields, especially immigrant women, because they are disempowered. All I’m saying is it’s a problem that exists in math just as it exists everywhere else too. It’s a real problem.
One of the reasons I left my job was I was asked to do stuff that the men didn’t want to do, and I said fuck that, I’d rather go work at a company where I get rewarded for my organizational abilities rather than exploited. I really think when people talk about women in STEM, they ask “Why don’t women stay in STEM?”, and I wanna say, “Hey, why don’t you ask why doesn’t STEM keep its women?” STEM is the one that needs to adjust; women are leaving because they’re brilliant and wonderful people and they have options. Why should you stick around at a job that isn’t treating you well? Fuck that! It’s STEM’s problem, not women’s problem.
” I really think when people talk about women in STEM, they ask “Why don’t women stay in STEM?”, and I wanna say, “Hey, why don’t you ask why doesn’t STEM keep its women?” “
You decided you wanted to leave finance and join the Occupy movement. What did it for you? Was it a single moment or was it a string of events that caused this?
I was disgusted by what finance was actually accomplishing, and I didn’t want to exploit the world. I didn’t wanna scavenge on the carcass of mistakes made by people that don’t know better. I left finance because I simply didn’t see how to have a positive effect within it. It was a string of events- it took me three years to figure out that we weren’t actually helping here. [For people doing finance internships,] you have to be an anthropologist: take notes, ask dumb questions like why does that help the economy? What are we really selling with that product? You might not be able to get answers but ask questions.
What was that like? And how did other people react to your decision?
I’m actually really good at quitting jobs. People say don’t burn your bridges, but my middle name is burning bridges. It’s very satisfying. They don’t like me, and I don’t care. The number one thing is to have a thick skin. I do think that if I had a thick skin earlier on I probably would’ve dealt differently with my colleagues in the math department.
Did you ever doubt what you were doing? At any point did you feel pressured to fall back? And how did you push through?
I have no idea what I’m doing. When I was 14 I decided to become a mathematician and it took me twenty years to become a math professor. That’s twenty years, and ever since then I’ve had at most a one-year plan; I’m risk loving. A lot of people say they are risk-loving, especially men in finance, but they’re lying; they never leave their job because they’re afraid. I can’t say that I was risk-loving without a fall-back plan, though.
I always advocate for ideas that I believe in, I don’t advocate for myself. I’m advocating for a concept of justice that I have thought about for a long time, and I’ve considered the push back for it already so I over-prepare. But once I’ve made my mind up I stick to it, but again the reason I can be so stubborn in my views is because I do it for the sake of the idea. Some of the best advice I’ve ever gotten was “Do what you would do if you didn’t have any insecurities”, but then I generalize that to “Say what you would say if you had no doubts.” Sometimes you’re wrong, but you learn faster that way. If you say it with a clear conscience that you’re trying to do your best and you believe in this thing then people can correct you and you change your opinion accordingly. It’s the fear of being wrong that keeps people from learning.
” Some of the best advice I’ve ever gotten was “Do what you would do if you didn’t have any insecurities”, but then I generalize that to “Say what you would say if you had no doubts.” “
You propose in your book that to alleviate the damage that WMDs cause data scientists should include human values in their models, and even suggest that we may have to forgo some accuracy. While this is a good idea, realistically a lot of the people making these WMDs are not aware of the social issues around them or have never though about unpacking those notions. What do you suggest we need to do in order to adjust the way they think?
First of all, there’s a lot of questions that the modelers have just not bothered to ask themselves, so the fact that there’s no current tools to fix some of these problems right now doesn’t mean that there won’t be, it’s just that they don’t care. We have to make them care. There was a New York Times piece by Joy Buomlawini about how facial recognition software worked 99% well on white male faces and only 65% on black female faces, and the way I see it is why didn’t they check that? Why isn’t it just a standard operating procedure when you’re writing code to show it’s accurate and inclusive? There’s no standards because they don’t care, but once they care they’ll be able to fix that. They’re smart, they’ll figure it out.
Big data benefits corporations not only time-wise but also profit-wise, so do you think it’s realistic to assume there will be a reform in the models that are currently implemented? Are we doomed?
We’re doomed right now, but I don’t think we have to be doomed. I think your generation is gonna make it better. Your generation grew up with a financial crisis and started caring about social justice way younger and more seriously than my generation. You guys are much more diverse and you have more of an urgency about how obviously shitty things are right now.
On your website’s About page you say you wish to some day answer the question: what can a non-academic mathematician do that makes the world a better place? Can you attempt an answer? Or do you have tips?
I’m starting a company to audit algorithms. I don’t have the leverage to get companies who are probably really bad actors to admit it, but I am slowly getting more and more clients so my idea is to get algorithms to show the flaws.
I would say the average person needs to stop being afraid of math. Actually, they’re allowed to be afraid of math if they want, but they can’t be afraid of algorithms. Algorithms are just secret decision making processes and they can be used very unfairly, so what I’m trying to promote is this idea that everyone gets to ask questions about how they got a job or didn’t get a job, about whether they qualified for a credit card or a mortgage or insurance. Those are decisions that you should understand, it’s a social justice issue. We need to understand whether they’re racist or classist and how they work. It’s not about math, it’s about justice, and everyone understands justice.
What advice do you have for girls and young women who want to pursue math?
The first thing I would say math is very patient, so if you worry about not being really really fast don’t worry about it. Math doesn’t mind if it takes you a while. We have set up a way to learn math in this society to make it seem like everything needs to be fast, but there’s really no reason.
The second thing is that math can be really beautiful in its own right and that could be why you’re interested in doing it, but it also can be very applicable. I wrote another book called Doing Data Science which is a nice introduction for how to use basic math techniques and basic programming and basic statistics to become a data scientist and to answer really interesting questions with data while being careful that you’re not propagating bias.