MIT Technology Review Narrated cover image

In Machines We Trust: The AI in the newsroom

MIT Technology Review Narrated

00:00

Large Language Models Aren't Training to Output the Truth

Large language models are trained on huge vast data sets collected from all over the internet. But as a result of this, there are few protections for people whose data they're actually using. For large language models like GPT, it could be a phone number or an email address. There's also the potential for misinformation and spam because large language models aren't trained to output the truth.

Transcript
Play full episode

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
Get the app