Chat GPT Understands & Reuben Cohn- Gordon

Chat GPT, an AI powered chat-bot, has become the world’s fastest growing application, with over 100 million users in the first month of its launch. Even its harshest critics concede that when interacting with Chat GPT, it can seem as if one is talking to an intelligent machine. But, the standard critique goes, that’s just an illusion. Chat GPT isn’t in fact intelligent. It doesn’t understand the questions it’s asked, or the answers it gives. 

But, what if this critique is wrong? What if our elevation of human understanding to something that machines like Chat GPT can’t reach is mere narcissism, or worse, a philosophical mistake? What if, what current AI can do isn’t really possible without some robust level of understanding? 

Reuben Cohn- Gordon is an AI researcher at the Iniversity of British Columbia and UC Berkeley, and recently wrote an article entitled GPT’s very Inhuman Mind for Noema magazine, in which he argues against the standard critique of large language models like Chat GPT, namely that they lack any form of intelligence and understanding. Reuben uses a number of ideas from 20th century philosophy in approaching this, as well as intriguing metaphors from Classic Roman literature. 

If you enjoyed the episode, please leave us a rating and a review on Apple Podcasts.

This podcast is created in partnership with The Philosopher, the UK’s longest running public philosophy journalm founded in 1923. Check out the latest issue of The Philosopher and its online events series: https://www.thephilosopher1923.org

Artwork by Nick Halliday

Music by Rowan Mcilvride