How did your journey in AI start - what brought you to AI?
I started taking courses on AI because I wanted to see what all the hype was about. But then I lost some confidence in my skills because I had struggled with advanced calculus before, and decided not to pursue AI because I did not want to go into a field if I could not become the best at it. But when I was presenting someone else’s research on gender bias in AI at the KAIST Conference on AI Fairness and Gender, I realized that I know a lot about this topic. The technical professor that was presenting seemed to not grasp the social issues well, whereas the social scientists in the audience seemed confused about the technical aspects. I realized there is a gap and because of my unique technical background and experience with gender equality work, I could be that bridge.
What is the biggest change you have observed in AI compared to when you entered the field at the start of your career?
When I first started writing my master’s thesis on gender bias in AI in 2020, there was very little research on this topic and I was struggling to find papers on this specific topic for my thesis. In the last year, a lot has happened and a lot of funding and initiatives have become available. The demand for AI ethics competency is growing and will explode once the EU AI Act regulations are implemented in a few years.
What excites you about working in AI, and what are you working on right now?
The impact of an AI app can affect the 5 billion mobile users in the world at the click of a button. What excites me is working with developers and researchers in AI to find out how to make sure this impact is net positive.
Right now, I am working on building my AI ethics consultancy business, Bui Consulting. I am trying to figure out what the most impactful thing I could use the next few years of my career on.
How do you see the number of female researchers and professionals in AI grow? How do you involve yourself in initiatives to introduce more women to AI, and what initiatives can we make to inspire more women to pursue careers within AI?
So far, I have not seen any signs or numbers that the number of women in AI to be increasing. On the contrary, the progress is going backwards. West et al. (2019, p. 10) report that the trend in both the professional field and computer science education in the USA are going backwards; numbers from 2013 and 2015 are lower than compared to the 1960s and 1980s, respectively.
A book that was published in 1983, wrote that the field of AI had an "unusually large number of women" compared to the rest of the field of computer science, whereas today, the field of AI has a lower amount of women than computer science in general (Forsythe, 2001, p. 165). Weissman estimates that the whole field of AI only had 13.5% women in 2016, which is lower than the number of women in ICT in the same year; The number of women in ICT in Europe (16%) , Americas (22%), and Asia (26%) were all higher in 2016 (as cited in United Nations University & EQUALS, 2019, pp. 85, 96).
It is important to understand that the diversity crisis is not a pipeline issue but rather a cultural problem. “Fixing the pipeline” is the approach we have had the last few decades, and it is not working because it puts the blame on women and downplays the cultural issues. When women with a computer science education are 14% less likely than men to work in STEM (Sassler, Michelmore, & Smith, as cited in United Nations University & EQUALS, 2019, p. 84), and nearly half of the women who enter technology eventually leave the field, which is twice as much as the percentage of men who leave the field (Ashcraft, McLain, & Eger, 2016).
It is time for the field to address issues like sexual harassment, wage gaps, and the abuse of power imbalances. I think it is important to openly talk about these issues and implement structural changes that keep perpetrators accountable and prevent them from doing it in the first place.
Why is it important for more women to get involved in AI?
Because gender bias in AI is a massive invisible problem that is oppressing women from success and is can undo decades of feminist progress. When Amazon created a resume-filtering AI, it sorted out all of the women. Using an AI like that would undo all progress in increasing diversity in AI.
What advice would you like to give women who are pursuing their careers in AI?
You don’t have to be the best to have something valuable to bring to the table. Your lived experiences and the things you care about because of the type of problems you encounter is a great contribution. Who is going to create AIs that predict productivity levels during menstrual cycles if male developers tend to solve their own problems when they do a startup? Who is going to solve the issues of sexist algorithms? I do not think women or other marginalized genders can wait for cisgender men to solve our problems.
What advice do you have for organizations who wish to recruit and retain female researchers and professionals in AI?
Do not address it as a pipeline problem but rather a cultural problem. Dare to talk about, track, and address taboo issues like sexual harassment and wage gap.
Who is your biggest role model within your field, and why?
My role model is Timnit Gebru who cared more about the integrity of her research, what is right, and risked her career, when taking a stance against Google when they wanted to muffle her research.
What role can NORA take on to empower diversity and inclusion in AI in Norway?
NORA can put together best practice guidelines for member organizations in collaboration with diversity experts, and NORA can shed light on important issues so that they are addressed.
References:
Ashcraft, C., McLain, B., & Eger, E. (2016). Women in Tech: The Facts (p. 76). National Center for Women & Information Technology.
Forsythe, D. E. (2001). Studying Those Who Study Us: An Anthropologist in the World of Artificial Intelligence. Stanford: Stanford University Press. Retrieved from http://www.sup.org/books/title/?id=1342
United Nations University & EQUALS. (2019). Taking Stock: Data and Evidence on Gender Equality in Digital Access, Skills, and Leadership. Macau: EQUALS Global Partnership. Retrieved from EQUALS Global Partnership website: https://2b37021f-0f4a-4640-8352- 0a3c1b7c2aab.filesusr.com/ugd/04bfff_145a18e6425e47a1b90d0440f7476d0f.pdf
West, S. M., Whittaker, M., & Crawford, K. (2019). Discriminating Systems: Gender, Race and Power in AI (p. 33). AI Now Institute. Retrieved from AI Now Institute website: https://ainowinstitute.org/discriminatingsystems.html
About Cathrine:
Cathrine Bui specializes in paving algorithmic paths to both profit and justice.
What does profit have to do with AI justice? When Apple and Goldman Sachs released the Apple Card, their AI gave women 10-20x lower credit limits than their husbands - even when the women had higher salaries and better credit scores. Biased data creates noise in algorithms which is bad ethics, bad for business, and bad PR.
Cathrine combines her past experience leading the EQUALS Research Coalition on digital gender equality and technical background in computer science to bridge the gap between tech engineers and social science for more Responsible AI. Her University of Oslo master’s thesis on «Gender Bias in AI» won the Best Scientific Poster Award at the 2021 NORA NordicAIMeet and she also won the NORA Diversity in AI Award for her work in responsible AI.
Her company, Bui Consulting, provides audits, keynotes, workshops, management consulting, and AI development services - for your profit and a better world. The company was a part of the Google Veksthus incubator program and her past customers include companies like Cisco and DNB Bank.