Anthropic, OpenAI Back AI Safety Checks as Researcher Warning Spurs Slower Development
Updated
Updated · Los Angeles Times · Sep 16
Anthropic, OpenAI Back AI Safety Checks as Researcher Warning Spurs Slower Development
3 articles · Updated · Los Angeles Times · Sep 16
Summary
Anthropic said it will give independent outside evaluators ongoing, employee-like access to monitor safety practices, and OpenAI said it will adopt the same measure.
The rare alignment followed an Anthropic researcher’s resignation and warning that AI poses threats to humanity, prompting Dario Amodei and Sam Altman to argue development should be paced rather than halted.
Amodei’s broader plan calls for common safety standards, government-backed coordination among frontier labs and possible limits on dangerous uses, model releases and eventually self-improving systems.
Implementation remains uncertain because U.S.-China competition, IPO and profit pressures, questions over auditor independence and President Donald Trump’s light-touch approach all work against a coordinated slowdown.