Anthropic's AI Safety Researcher Now Briefing the White House

Nicholas Carlini, a researcher at Anthropic, has made a sharp pivot. Back in March, he raised alarms about the dangers posed by Mythos. Now he's part of an Anthropic team sitting down with the White House to discuss AI safeguards.

The twist? Carlini and his team are reportedly arguing that the latest AI models should be released. That's a notable shift from sounding the alarm to making the case for deployment — suggesting Anthropic believes its safety measures are strong enough to justify pushing forward.

The Wall Street Journal profiled Carlini's trajectory, highlighting how quickly the AI safety conversation is evolving. Researchers who flag risks one quarter end up advising the highest levels of government the next.

It's a sign of how central Anthropic has become in shaping the policy debate around frontier AI models.