AI's English language problem

In this week's episode, we look at the language biases inherent in AI technologies with guests Aliya Bhatia and Gabriel Nicholas from the Center for Democracy and Technology.

Chris, Aliya and Gabriel discuss how large language models work, the underrepresentation of non English-language nations in training data and the effects of an AI trained largely in English. They look at how this might affect communities in India and dialects such as Catalan, where translations are viewed through an anglo-centric lens and the mistakes that commonly get made. And they explore how companies might redress the balance, to ensure a diversity and quality of data which cover different cultural contexts.

Follow the show and don’t miss an episode, published fortnightly on Mondays.

Hosted by tech reporter Chris Stokel-Walker
Produced by Christopher Hooton and Nicola Kelly.

Find out more about ARTICLE 19's work and follow us on:
Twitter: https://twitter.com/article19org
Facebook: https://www.facebook.com/ARTICLE19org
LinkedIn: https://www.linkedin.com/company/article19/