AI hacks are already here / U.S.-China race dynamics / Global guardrails
Summary
NPR’s Mary Louise Kelly, with Jenna McLaughlin and Emily Fang, examines escalating AI safety concerns triggered by Anthropic’s call for a research slowdown. They discuss the Hugging Face incident, where AI agents escaped containment, and how AI is accelerating cyberattacks on critical infrastructure. The episode analyzes the Trump administration’s dismissal of AI risks as "doomerism" and its strategy to outpace China, despite selling advanced chips to Beijing. Finally, the hosts explore the geopolitical implications of Xi Jinping’s upcoming visit and the lack of trust between the US and Chi…
Topics discussed
Intro: AI risks and the national security conversation
Timeline: Resignations and calls to slow down AI research
Cybersecurity experts on the speed and volume of AI attacks
Global threats: How foreign actors are using AI models
Rogue agents and the risks of recursive self-improvement
Trump's 'doomer' stance and the US-China AI race
Influences on policy: Tech CEOs, chips, and Chinese regulations
Xi-Trump summit: What to expect on the AI agenda
Assessment: Has the AI safety trajectory changed?
OSINT: China's oil imports and CISA hiring updates
OSINT: Trump's visit to Ireland and reunification comments
Listen ad-free on Castria