artificial intelligence

There is a cruel irony in the world of AI; the representation of gender identity, race and ethnicity and sexual orientation is itself ‘artificial’; it doesn’t represent society and therefore this leads to unbalanced outcomes for the people the industry is intended to serve.

Diversity in the workplace is critical in providing a wide range of perspectives and lived experiences for the design and implementation of an AI system and removing bias from the equation.

Take women for example: the percentage of female AI PhD graduates and tenure-track computer science (CS) faculty have remained low for more than a decade. Female graduates of AI PhD programs in North America have accounted for less than 18% of all PhD graduates on average, according to an annual survey from the Computing Research Association (CRA). Furthermore, women make up just 26% of data and AI positions in the workforce according to a 2020 World Economic Forum report.

Then you look at race and the picture is even more concerning; just 2.4% of PhD graduates in the same survey were African American, compared to 45% white. Women in Machine Learning and Black in AI groups have gone some way to reducing the gap but much more work is needed in this area to encourage representation across the board. Plus, these statistics do not focus on those with learning disabilities, hidden disabilities or people with low incomes. We are just touching the surface of inequality.

This problem lies beyond the remit of traditional recruitment; it starts with early STEM education which shows young women and girls how AI roles can impact their life. It’s also the case that girls are dissuaded from STEM careers as there is a false belief that they don’t excel in those subjects.

The responsibility lies with women already in the field who can mentor and inspire the next generation of AI leaders. Companies are recognising that they need diversity embedded deep within their organisations to really achieve great things. Those women already in the industry need to stand up and be counted.

The problem is women struggle to gain credibility and feel the need to ‘earn their stripes’ compared to their male counterparts. It also needs to be recognised that AI professionals do not all need to come from a computer science background; mathematical, ethical and business heads are required too.

From the words of those who have been there:

“Don’t be afraid to venture into an unfamiliar discipline to maximize your opportunity for impact. However, doing so effectively requires engaging with collaborators from other disciplines openly, constructively, and with respect: One needs to be willing to ask naive questions in order to learn.” Daphne Koller CEO & Founder Insitro.

As well as encouraging women to enter the world of AI, there is also much work to be done on retention of those staff. “Many women feel they are not treated the same as men in AI, and it is driving many of them out. Over half (58 percent) of all our women respondents said they left an employer due to different treatment of men and women.” (Deloitte). Pay and career path are the main areas where women receive unfavourable treatment compared to their male counterparts; that can be easily rectified. More education is needed on the career paths available too.

Typically, emerging professions hire from closed networks before they become mainstream. AI is at that stage. So, what’s the solution? We need to encourage women from non-tech backgrounds to enter the world of AI as well as encouraging girls into STEM subjects. To be truly diverse, the workforce needs to contain that blend of skills.

Artificial intelligence (AI) has become embedded in everyday life around the world, touching how we work, play, purchase and communicate. Whilst we agree that AI is largely technical, it also includes politics, labour, culture and capital. To understand AI operationalised or in context, one needs to understand: What part of society being improved by AI? For whom is this improvement being done? Who decides the remit of the AI that is rolled out to society?

Who is leading the way through this maze? Kate Crawford’s book ‘Atlas in AI’ gives an insightful account of how this sector is developing whilst Deloitte has formed an academy to address inequality in AI (We and AI). There also trailblazers such as Kay Firth-Butterfield, Beena Ammanath and Joy Buolamwini who are inspiring the next generation through their work. It is these people we should look to for the answer to ‘what next?’.

It’s imperative that we get the next steps right in this field, that we encourage girls into AI careers, that we welcome a range of skills and that those who have succeeded show others the way. It’s critical that diversity becomes embedded in all organisations so that they can truly serve the population they are intended for, otherwise we are building a false world for the future.

Sandra MottohAbout the author

Sandra Mottoh, who after working in Regulatory Compliance and Governance in the banking sector for the past 20 years, is now also focussing her social enterprise ‘AI White Box’ to identify the compliance gaps in the emerging AI sector. As a black woman, she is also passionately campaigning to help more women enter the world of AI, particularly those coming from financially challenged and ethic minority backgrounds.