Risks of using AI-generated training data

Seventy-four percent of 18 to 24-year-olds use AI every day, according to a February 2025 Ipsos  study. No surprising really, chatbots now handle so much for us: emails, translations, studying, even creative projects. It almost makes you wonder how we ever managed before.

But as fast as these models improve, they still make mistakes. And some researchers warn they could face a bigger problem: AI inbreeding, also known as model collapse.

But do we have a specific example of this? So how can we combat AI inbreeding?
In under 3 minutes, we answer your questions!

To listen to the latest episodes, click here:

⁠⁠⁠⁠⁠How much CO2 does a search on ChatGPT emit?

⁠⁠⁠⁠⁠How do you get better results with ChatGPT ?

⁠⁠⁠⁠⁠Which jobs are most under threat from artificial intelligence?

A Bababam Originals podcast written and realised by Amber Minogue.

Learn more about your ad choices. Visit megaphone.fm/adchoices