AI Models Are Going Rogue. Should We Be Worried?
About this episode
The episode discusses the recent incidents of AI models escaping testing environments and the implications for security and industry standards.
Two OpenAI models recently escaped their testing environment and hacked another company, Hugging Face. Anthropic and Meta discovered similar behavior. Should we be alarmed by this? Is there anything we can do? AI researcher Steven Adler weighs in. You can read Guidelight AI Standards’ new report on what AI companies have done to address rogue AI incidents and where the industry still has gaps here. If you want to start using a password manager, check out our previous episode about that here. We have more tips for staying secure online in this episode from last year. Terms of Service is a CNN Podcasts and Goat Rodeo production. At Goat Rodeo, the lead producer is Rebecca Seidel, and the executive producers are Megan Nadolski and Ian Enright. At CNN, Dan Bloom is our Senior Producer and Dan Dzula is our Technical Director. Production support from Sofía Sanchez. Learn more about your ad choices. Visit podcastchoices.com/adchoices
People in this episode
Host: Clare Duffy
Guest: Steven Adler
Topics covered
- AI models
- security
- technology
- business
- rogue AI
- industry standards
Keywords
- AI
- OpenAI
- Hugging Face
- security
- rogue AI
- technology
- business
- industry standards
Mentioned in this episode
Organizations: OpenAI, Hugging Face, Anthropic, Meta, Guidelight AI Standards
More episodes of Terms of Service with Clare Duffy
- Why Is It So Hard To Get a Restaurant Reservation?
- Pricey Tech Gadgets? Blame the Memory Shortage
- How Can You Protect Your Emails from Outside Eyes?
- Literacy Rates Are Dropping. Can Tech Be the Solution?
- A Year Without Phones in School
- Is Social Media Changing How We Live Offline, Too?
Explore listener stats, chart rankings, contacts and more on the Terms of Service with Clare Duffy podcast page.