As AI works its way into more parts of everyday life, the biases it exhibits are coming under closer scrutiny. Canvas8 spoke to Robert Elliott Smith, an AI expert and senior research fellow at UCL, to find out how developing diverse artificial intelligence can help businesses and society alike.
“We live in a world increasingly ruled by technology; we seem as governed by technology as we do by laws and regulations. Frighteningly often, the influence of technology in and on our lives goes completely unchallenged by citizens and governments. We comfort ourselves with the soothing refrain that technology has no morals and can display no prejudice, and it's only the users of technology who distort certain aspects of it.”
This isn’t the promo for the latest dystopian sci-fi blockbuster, it’s the blurb from Rage Inside the Machine, the 2019 book by Robert Elliott Smith, an AI expert and senior research fellow in computer science at University College London. It paints an accurate image of AI becoming increasingly intertwined with everyday life. “This spans from how we select our mates through dating apps to how we [find] news, and from how we recommend probation for people who have been in jail to how we recommend medical outcomes,” says Smith. In fact, research carried out in 2016 found that two-thirds of people already trust AI to handle medication reminders, travel directions, entertainment, targeted news, and manual labour.
A lack of diversity in data sets can lead to prejudiced AIs
TED Conference (2019) ©
With just 16% of global internet users saying they find brands through personalised product recommendations, there’s an opportunity for brands to offer more varied viewing and shopping tips. People want these recommendations to be based on more than just past purchases ‒ 32% of UK and US consumers want suggestions based on their budget and 21% want recommendations based on their personal style. Tapping into this, online styling service Stitch Fix sends users boxes of clothes created by cross-referencing their stated preferences with the recent purchases of others who fit similar demographic profiles. Meanwhile, HBO’s ‘Recommended By Humans’ platform helps people find something new to watch by providing guidance from real people and social media.
To read the full report on the Canvas8 library, click here.
Robert Elliott Smith is the chief technology officer at BOXARR Ltd, a senior research fellow in computer science at University College London, and a founding member of The UCL Centre for The Study of Decision-Making Uncertainty. His research focuses on complex systems-based artificial intelligence and he has authored over 20 journal articles, 50 conference papers and conducted research projects for companies including the US Army Strategic Defense Command, The Center for Nonlinear Studies, NASA, and Boeing.
Rebecca Smith is a senior behavioural analyst at Canvas8. She has worked with global clients including Google for Education, Dentsu Aegis, and Superbrands. Outside of work, you’ll find her searching for the UK’s best cacio e pepe or with her nose in a fantasy novel.
13 Aug 19
2 min read