White House & Anthropic Team Up on AI Safety

0:00It's April 17th. You're listening to U.S. News Today, AI-powered coverage from across America. I'm Cory with the story. White House reps just sat down with anthropic bosses for what they called a productive talk on teaming up for AI, stressing that balance between fresh, innovation, and keeping things safe. This comes after serious beef between the Trump team and anthropic, who push hard for guardrails on AI to cut risks in boost U.S. security. Trump tried banning federal use of their clawed chatbot over a pentagon contract fight back in February, and Secretary Pete Hexeth tagged them as a supply chain threat, but courts like Judge Rita Lynn's March ruling shut that down. Reactions are mixed, but leaning impressed, and anthropic critic David Sacks, ex-White House AI's R, says Mythos. Their April 7th model that outs Mark's humans at spotting cyberholes is no hype. Its real progress as coding AI's gets sharper at exploits. Anthropics limiting Mythos to top clients and kicked off project glasswing with giants

1:04like Amazon, Apple, Google, Microsoft, and JP Morgan to hunt. Software bugs safely. The UK's AI Security Institute tested it too, calling it a clear step up with warnings on weak systems. Co-founder Jack Clark warns more beasts like this are dropping soon from rivals and even open models out of China so that White House vibe on CoLab feels timely to steer the power without the pitfalls.