John MacCormick, Dickinson College – Fairness First in Artificial Intelligence

A.I. is not without bias.

John MacCormick, professor of computer science at Dickinson College, discovers the thriving research field of fairness in A.I.

John MacCormick is the author of three books, including Nine Algorithms That Changed the Future: The Ingenious Ideas That Drive Today’s Computers and What Can Be Computed?: A Practical Guide to the Theory of Computation. Dr. MacCormick holds 19 US patents on novel computer technologies and is the author of numerous peer-reviewed articles; his Nine Algorithms book has been translated into eight languages.

Fairness First in Artificial Intelligence

In the last 12 months, we’ve seen some tectonic shifts in the landscape of artificial intelligence. Chatbots such as ChatGPT respond to our questions with unprecedented fluency, and image generators such as Dall-E 2 and Stable Diffusion produce novel images based only on a text description. We’re only just beginning to glimpse where these new technologies will take us. But there’s one thing we already know about this new world of artificial intelligence algorithms: they can reflect, and sometimes amplify, the unfairness that already exists in our society. For example, suppose you ask the Dall-E 2 image generator for a picture of a real estate broker. In the US, more than half of real estate brokers are women. But the chance that you generated a picture of a woman is only about 30%. So women are systematically underrepresented in the results. And there are similar problems with the race and ethnicity of people in these generated pictures. That’s why researchers are working to reduce bias in artificial intelligence algorithms. Fairness in AI is in fact a thriving field of research these days, with its own conferences and academic journals. But algorithmic bias is an incredibly hard problem to solve. If AI is to make the world a better place for all of us, then academic researchers and technology workers will need to focus even more on how to achieve fairness in artificial intelligence.

Read More:
I unintentionally created a biased AI algorithm 25 years ago – tech companies are still making the same mistake (theconversation.com)

Nine Algorithms That Changed the Future | Princeton University Press

What Can Be Computed?