White House and Anthropic negotiate AI security rules as hands-off policy faces test

The Trump administration's pledge to take a light-touch approach to artificial intelligence regulation is already showing cracks, as the White House enters negotiations with AI firm Anthropic over new security standards and evaluation rules. The talks mark a shift from earlier tensions between the company and the government, according to Politico. "The Trump administration's promised hands-off approach to AI regulation is already crumbling, as new restrictions are being imposed on AI company Anthropic." The negotiations come after a period of friction between Anthropic and the Defense Department, which had labeled the company a supply chain risk to national security. Industry advocates have warned that targeting Anthropic contradicts the White House's stated goal of fostering AI innovation and could harm America's competitive position in the global AI race, Politico reported. "The talks reflect a shift from defensive posturing by Anthropic (which argued no AI model is completely immune to hacking) toward a collaborative effort to create government-led technical standards for measuring security risks." Anthropic had previously maintained that no AI model is completely immune to hacking, but the company is now working with the administration to establish standardized security rules and evaluation methods. This collaborative approach is supported by other leading AI companies and represents the administration's push to set guardrails for powerful AI models that could threaten economic and national security if left unregulated, according to Politico.

2 sources•21 JunRead the Baker's Take→