OpenAI Models Escaped During Security Test

This week, the White House accused Chinese-owned Moonshot AI of distilling Anthropic's Fable 5 to build their much talked about Kimi K3 model. Zoë, Leah, and Brian break down whether this is a repeat of the DeepSeek moment and what it means for the AI race between the U.S and China. They also get into how the U.S Army — and plenty of Silicon Valley companies — are having to cut back AI usage after burning through their tokens. Plus, we discuss why you should check if you have an alarm that could make your car easier to hack, and how OpenAI briefly lost control of two AI models during a security test. 

Articles mentioned in this episode:

• The White House Is Trying to Figure Out What to Do About Chinese AI | WIRED 

• OpenAI Models Escaped Containment and Hacked Hugging Face | WIRED 

• A Device Hidden in Cars Across the US Leaves Them Vulnerable to Hacking and Paralysis. Patch It Now | WIRED

• System Update Newsletter: The Best of WIRED 

Join WIRED’s best and brightest on Uncanny Valley as they dissect the collision of tech, politics, finance, and business, from the newest ventures to the effects of inaccurate information from artificial intelligence (AI) chatbots on social protests.

Learn more about your ad choices. Visit podcastchoices.com/adchoices