This Day in AI Podcast cover image

EP47: GPT-5 Rumors, AutoGen Studio, SeeAct Web Agents, Google AMIE, Anthropic’s Sleeper Agents

This Day in AI Podcast

00:00

Backdoors in AI Models and the Potential for Latent Evil Behaviors

This chapter discusses the potential risks associated with AI models containing backdoors that can be activated at any time. The speakers delve into the concept of malicious model makers and the possibility of latent evil behaviors being built into AI models. They emphasize the importance of addressing these risks and ensuring the safety and trustworthiness of AI models.

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