ChatGPT isn’t as smart as we think

Artificial Intelligence and chatbots are having a mainstream moment. In November, the public was introduced to ChatGPT – a chatbot that can have seemingly human-like conversations with users. And after a “creepy” conversation between a New York Times tech columnist and Microsoft’s new Bing chatbot (which called itself “Sydney”) the debate around AI sentience has re-ignited.

But, behind all the awe, argues AI researcher, author and data journalist, Meredith Broussard, is a model that’s simply really good at math – and the technology that powers our AI today can often be biased, sexist and racist. She’s on the show to talk about how we should all be thinking about these problems in a tech innovation that isn’t going away anytime soon.

Questions? Comments? Ideas? Email us at thedecibel@globeandmail.com