Chloé-Agathe Azencott

In France, the relative proportion of women among scientific and technical students is increasing — but not in the most prestigious programs in mathematics and computer science. In 2016, fewer than 10% of computer science engineering students in France were women. Villani's report on artificial intelligence reminds us that 12% of professionals of the digital industry holding technical jobs are women. The situation isn't much better in the rest of the world. Those numbers are worrying: first, we must wonder what it is about these studies and jobs that repel women; and second, we must worry about the biases the lack of diversity induces in artificial intelligence applications.

I am committed to promoting diversity and inclusion in science, technology, engineering and mathematics (STEM). In addition to social justice considerations, I strongly believe that diversity, be it of genders, sexual orientations, ethnic origins, socio-economic backgrounds, family status, religious beliefs, or physical abilities, opens us to valuable life experiences and increases our ability to ask important questions and solve problems. I also believe this is particularly important in both machine learning and the biomedical sciences, where the perspective of underrepresented people is key.

There is so much I would like to do to promote diversity in STEM, but one has to start somewhere. For me, this therefore primarily means two things:

• Fighting unconscious biases when recruiting.
• Promoting women in machine learning and data science, through my visibility, through outreach activities targeted towards high-school students, and through Paris Women in Machine Learning and Data Science. Paris WiMLDS is the Parisian chapter of WiMLDS. We host regular meetups where all invited speakers identify as female. People of every gender are welcome to attend! We also have a public Slack channel.

Women-in-<please fill in the blank> events

I am usually happy to be invited to talk about my research and my career at any women-in-<please fill in the blank> event. I'm also happy to talk specifically about women-in-<please fill in the blank> issues, but I'm going to get angry if all the women speaking at your event are on the women-in-<please fill in the blank> panel and none or very few of them are giving technical talks; although I'm knowledgeable about these issues and I believe my experience to be valuable, you may want to consider inviting someone who actually studies women-in-<please fill in the blank> instead, and I can tell you about machine learning, precision medicine, and other things I'm actually studying.

Are you trying to organize a gender-balanced machine learning event, but find yourself unable to find women speakers? WiML has a great list of women active in machine learning. You may also be interested in Request a Woman Scientist and, for French speakers, in Les Expertes.

You will find below a few references on diversity in STEM.

Unconscious biases

• Northwestern's resources on unconscious bias.
• The University of Arizona's Avoiding gender bias in reference writing (pdf).

International reports and recommendations

• I'd blush if I could: closing gender divides in digital skills through education by UNESCO (May 2019), with an important chapter on female voice assistants.

Barriers against women in STEM

• A very thorough and thoroughly referenced text on Sexism in the Academy by Troy Vettese (published in 2019).

• A study on persistent gender biases in PECS (physics, engineering, and computer science), showing that interventions to close the gender gap may work to attract high-achieving women, but that there are still factors at play that attract low-achieving men and repelling average- and low-achieving women (published in 2020).

• Even when women become well represented, gender biases persist (published in 2020).

• Women are perceived as less able students than men even though they outperform them (published in 2020)

• Gender Bias in Academe: An Annotated Bibliography of Important Recent Studies by Danica Savonick and Cathy N. Davidson (last updated May 2018, at the time I'm writing this).

• The Gender-Equality Paradox in Science, Technology, Engineering, and Mathematics Education (published in 2018)

• Women do ask for money. They just don't get it (published in 2019)

Here are many links to scientific studies on the topic, courtesy of Celeste Labedz on March 24, 2019, which I am archiving here in case Twitter loses them or something:

• Scientists were more likely to hire a hypothetical candidate for a lab manager if the name was male.
• Peer reviewers found women authors to be less competent, even when their work had similar impact levels.
• Women are less likely to be offered opportunities to speak about their research at top universities.
• Recommendation letters have pretty different content when written for men vs women, and it's not in women's favor.
• Female grant applicants are equally successful when peer reviewers assess the science, but not when they assess the scientist.
• There's not any evidence that women are less competent earlier in the "pipeline" when less exposure to bias has taken place.
• Women's service load is higher than that of men.
• Teaching load could put female scientists at career disadvantage.
• Gender and international diversity improves equity in peer review.
• Women in STEM are the frequent victims of inequity.
• Women are the targets of sexual harassment in STEM as well.
• "Meritocracy" is often synonym for "biased against women".
• Yes, representation matters.
• Men tend not to believe in any of theses studies.