Artificial intelligence and inclusive design
Maxine Mackintosh is a data scientist, founder of OneHealth Community, and one of those women who has the kind of innate charisma and deep intelligence that, I would go as far to say, leaves her entire audience in awe.
I saw her speak at an event on Women in AI, run by CognitionX in London, and was keen to hear more about her involvement in the Open Source community.
During her talk, she cited some intriguing stats about the gender imbalance within GitHub, and the potential that it has to influence the development of the next generation of AI technology.
GitHub has expanded its remit from being a functional cloud service that programmers use to store, share and work on their software projects, to a social network and online community. In the words of Open Source veteran Josh Wulf:
GitHub is a social network where your social capital, created by your commits and contribution to the global conversation in whatever technology you are working in, is yours — not tied to the company you happen to be working at temporarily.
To say it is of immense importance in the software development world would be an understatement.
However, there are issues in the GitHub community that Maxine highlights. Lack of gender diversity, for example, is a much discussed issue; a study in 2013 found that only around 11% of its base are women.
This is a concern, and it goes far beyond gender diversity
As GitHub further establishes itself as the heart of the Open Source community, with more and more developers collaborating on it, it is at the centre of the artificial intelligence software evolution.
This was the premise of Maxine’s talk and the focus of our conversation. She believes that the creators of artificial intelligence and other ground breaking technology should have a fair representation of the society it will then reside in.
Sadly, in the business world, there is an arms race to take advantage of this next wave of technology, and from what I can see, diversity in hiring is definitely not the top priority.
Making the case for inclusive technology
Just as we know the benefits of having a diverse selection of backgrounds on a board is good for the bottom line, what Maxine and other experts in the field of AI are suggesting is that diversity on teams is key if we want to build inclusive technology.
Sadly, this isn’t just about ensuring greater benefits. It’s also about protecting the less powerful groups in society.
We have already seen the dangers of building technological experiences that perpetuate damaging privilege dynamics.
Airbnb offers a well-cited example; its creators didn’t account for racial prejudice entering the platform, and had to take immediate action when they found out hosts were preventing people from renting based solely on their race.
It begs the question: are we willing to stand by and watch while the next generation of technology companies are built by the same types of people, with the same diversity issues?
So how do we get there?
Maxine was not just calling out the problems; she also had several ideas as to how we can tackle this issue:
Rigid, strict quotas. This has to go beyond allowing companies to demonstrate their ‘best efforts’, because otherwise, we are not taking AI seriously enough.
Publicising examples of AI gone wrong
Maxine believes that when companies begin to realize the reputational risks they could face if their AI creations alienate members of their customer base, they will begin to take their responsibility more seriously.
The Microsoft bot, Tay, is a classic example — it quickly absorbed racist and sexist views from the humans interacting with it, and adapted to those views, becoming a misogynistic racist itself.
Encourage more women to join the open source community
Of course, this is a problem that many organizations and governments are trying to solve; how can we get more women into technology? Maxine mentioned that increased engagement from women when coding is less of an isolated activity, and involves in-person group sessions or learning circles.
If we are able to create an environment which encourages women to try coding, and gives them the confidence to contribute to communities such as GitHub, we may begin to see some progress in the diversity balance.
I want to thank Maxine for her time. It was such a pleasure to hear her thoughts on why a diverse, representative open source community is key for the next stage of development of AI.