The importance of representation among the creators and designers of artificial intelligence cannot be underestimated. “Right now, something like 80% of the people doing the design are men,” according to Audrey Mandela, chair and co-founder of Women in Telecoms and Technology.
This disparity gives rise to many problematic situations. Sarah Burnett, chair of BSC Women, pointed out that a lot of AI assistants, including Alexa, Siri and Cortana, are female, which might indicate that women are seen by AI designers as being in supporting, rather than leading positions. “Of course!” was the straightforwardview of Pat Chapman-Pincher, mentor and non-executive director, who described herself as “not a great optimist about AI.”
"There have been voice recognition systems which don’t recognise women or women of colour."
Mandela gave examples of AI being used in candidate screening for HR and in voice recognition. “When AI is being used to employ people, it looks at what success looks like and success looks like a white guy. There have also been voice recognition systems which don’t recognise women or women of colour. They only recognise white men’s voices,” she said.
To address this matter and avoid such grave flaws taking place in the future, it is imperative that those designing, training and teaching AI systems come from a diverse range of genders, ethnicities, cultures and classes.
When questioned by the audience as to what can be done to bring more women into the field of AI, Laetitia Cailleteau, managing director at Accenture UK and Ireland anddirector of Liquid Studio which hosted the panel, said that conscious decisions have to be made when hiring.
"We have to really force diversity because otherwise it just doesn’t happen.”
She admitted: “It’s hard work. It means that to find the same calibre of people, you may find yourself in a difficult position for a few more months. I do monitor diversity. We have to really force it because otherwise it just doesn’t happen.” Cailleteau added that, when someone is about to leave her team, she has been known to say, "I only want a woman replacing you."
Clara Durodie, founder and managing partner at Cognitive Finance Group, said that role models play an extremely important part in getting women to enter the AI industry, especially those who are looking to switch from their current careers. She said: “This visibility of the role models is something we need to be active about.” She explained that she is a fund manager by training and an economist by education. However, in the last three to five years she has pivoted her career by: “spending time in the AI space, learning how to code, rewiring my brain to think in a different way about this technology.”
“Through our work and visibility we can inspire younger women and women in the current workforce.”
Durodie said that women leaders in AI need to “put themselves out there” in the public sphere and especially in the media. She said: “Through our work and visibility we can inspire younger women and women in the current workforce to change their careers.” Burnett pointed out that this is starting to happen, stating that Dr Sue Black, one of the most influential women in technology in the UK and the founder of BCS Women, had recently been interviewed on BBC One’s The One Show.
Another member of the audience posed the question of what could be done in school classrooms to encourage girls to enter the field of AI.
Cailleteau said that at Accenture they are involved in initiatives, such as Hour of Code and Girls Who Code, which give children insight into what a developer does. She said that there is a role for the government to play also, in terms of pushing and centralising some of these initiatives in the curriculum.
Burnett had already told the audience during her introduction that the aim of BCS Women is to encourage more women into the IT industry in general. However, in response to this question, she detailed specific activities of the group. She said a couple of years ago BCS Women ran an “app-athon” where parents brought children along to experience coding in 300 centres across the country. The children were taught to program Android phones to make cat and duck noises.
She said: “Things like that really appealed to kids. We could see that at the end of the sessions, the kids were really inspired. We need to do the same with AI.”