Gender diversity in AI not improved in 25 years

Research from Nesta has found that less than 14% of papers published on AI have at least one female co-author, a figure not improved since the 1990s

A report released today, by innovation foundation Nesta, has revealed a definite lack of gender diversity in the field of artificial intelligence (AI).

The report, titled Gender Diversity in AI Research, and based on papers published by the well-cited repository ArXiv, shows that only 13.83% of AI publication authors are women, and that in relative terms, the proportion of female co-authors in AI has not improved since the 1990s.


Related news: Not enough being done on diversity in AI, finds BCS report


Location and research domain also have an impact on the proportion of women in AI publications. For instance, the UK is behind the Netherlands, Norway and Denmark in this metric, but ahead of Japan and Singapore. Coming in at 22nd on Nesta’s list of countries, 26.62% of AI papers from the UK have at least one female co-author.

Women working in physics, education, ethics and biology are also more likely to be named as co-authors on published AI papers than those in the computer science and mathematics sectors.

The report states: “We note than AI papers in computer science research domains such as machine learning, and data and informatics have a significantly lower probability of containing at least one female author after controlling for other factors.”

As the report makes clear, female researchers are helping to do more to analyse the social implications of technologies. AI and tech’s unique ethical challenges make encouraging more women into AI a vital step.
– Priya Lakhani, Century Tech

Advertisement

The gender diversity gap in universities is also clear. The University of Washington tops the list, with 25% of AI researchers being female, but it is the only university to achieve this score. Every other organisation in the dataset scores below 25%, with ETH Zurich coming in at the bottom of the list with only 10.15%.

Big tech shows a similarly bleak picture, with only 11.3% of Google employees having published AI research being women, and similar figures showing for fellow tech giants Microsoft (11.95%) and IBM (15.66%).

In an interview with Nesta, Eve Riskin, associate dean of diversity and access in the College of Engineering at the University of Washington, described the ‘toxic environment’ and ‘pervasive imposter syndrome’ experienced by much of the female faculty.


Related news: Gender Balance in Computing project receives £2.4m government funding


Priya Lakhani, founder and CEO of AI company Century Tech, told Education Technology: “This important report highlights that we all have more to do to promote AI, and tech more generally, as a field in which people from all backgrounds can thrive. We simply cannot afford to miss out on the exceptional pool of female talent.

“Not only does this expand the number of brilliant minds working on some of our most important challenges, but as the report makes clear, female researchers are helping to do more to analyse the social implications of technologies. AI and tech’s unique ethical challenges make encouraging more women into AI a vital step.”

The full report is available to view at www.media.nesta.org.uk/documents/Gender_Diversity_in_AI_Research.pdf

Advertisement