Search

Women in Data Science and AI

Using research to inform policy measures aimed at increasing equity in the data science and AI fields

Project aims

The fields of AI and data science have grown exponentially as the world is increasingly being built around automated systems and 'smart' machines. Yet the people whose work underpins this are far from representative of the society these systems are meant to serve.

As AI becomes ubiquitous in everyday life, the drive for diversity and inclusion in technology is of pressing concern. In particular, the persistent under-representation of women and marginalised groups in data science and AI, including gender data gaps, leads to the encoding and amplification of bias in technical products and algorithmic systems, creating harmful feedback loops.

It is crucial that we get ahead of this now, before flawed technologies become irreversibly integrated into the fabric of society. We address these complex challenges through a three-tiered approach:

Mapping the participation of women in data science and AI in the UK, and globally. Women make up half the UK population, but only 20% of AI and data professionals and 18% of users across the largest online global data science platforms. Our aim is to increase the number of women in these fields.
Investigating diversity and inclusion in online and physical workplace cultures. We explore the ‘chilly’ organisational cultures of tech companies and institutions that result in persistent structural divergence in women’s and men’s career trajectories.
Exploring how the gender gap shapes scientific knowledge and technological innovation. The lack of diversity in the AI sector results in harmful feedback loops of social bias being built into machine learning systems. We promote responsible, gender inclusive AI design.

A demonstration of gender bias when translating Hungarian to English on Google Translate

A hiring algorithm developed by Amazon was found to discriminate against female applicants

The Apple Card's algorithm was accused of being a 'sexist program' in 2019

An independent audit found that Facebook was withholding certain job ads from women because of their gender

Scientists found music recommendation algorithms had a biased feedback loop

According to a 2020 BBC Investigation, women with darker skin are more than twice as likely to be told their photos fails UK passport rules

Publications

Where are the women?

Based on an original data set and innovative methodology, this paper shows the extent of gender disparities in careers, education, jobs, retention, seniority, status and skills in the AI and data science fields

The Digital Revolution

This paper, as part of the UN Women Discussion series, outlines a conceptual framework for understanding the mutual shaping relationship between gender and technology

Current work

Diversity in the Tech Industry

In collaboration with Diversio, an AI-based diversity, equity and inclusion (DEI) platform, this research explores behavioural and structural inclusion challenges faced by women in technology companies.

We are using Diversio’s database and employee public review websites’ data to analyse, and therefore tackle, diversity and inclusion pain points (e.g., pay gap, leadership diversity) within the technology industry across Canada, United States and the United Kingdom.

The Gender Gap on Stack overflow

Stack Overflow, the world’s largest question-and-answer platform of programming knowledge, is a key site for establishing technical expertise and reputation. The overrepresentation of men among the site’s elite may therefore compound the barriers to women’s participation in computing.

Using historical posts, answers and comments from the platform’s archive, our research explores how gender mediates reputation and knowledge creation on the platform, with the aim of providing actionable recommendations for inclusivity in online forums.

Research toolkit

Diversity Dashboard

We developed a tool, the ‘Diversity Dashboard’, to offer insights into gender inclusion within online tech workplaces. The Dashboard was designed for use by technology companies, running internally on top of GitHub and Slack, to measure and highlight in real-time potential differences between the treatment of men and women.

To find out more about the tool please watch the demo video below and read the information sheet.

LinkedIn scraper

We developed a new and innovative tool for scraping LinkedIn's profiles to access information regarding individual’s job, skills and educational backgrounds in the data and AI fields. For more information, please reach out via our contact form.