Anthropic's leaked federal AI bill criticized by safety advocates as weaker than bipartisan FRONTIER Act
ai-safety anthropic
| Source: Techmeme | Original article
Anthropic's leaked federal AI bill faces broad criticism from AI safety advocates for being far weaker than the bipartisan FRONTIER Act, undermining its safety‑first claims.
Anthropic’s draft AI‑safety legislation, circulated among Capitol Hill staffers and leaked late last week, has ignited a sharp backlash from the AI‑safety community. The proposal, which the company has been promoting as part of its “safety‑first” brand, is markedly less stringent than the bipartisan FRONTIER Act that has been gaining traction in Congress. Critics say the draft dilutes key safeguards, making Anthropic’s public safety commitments appear hollow.
The leaked document, first reported by Politico and subsequently highlighted by Yahoo, the Washington Post and Forbes, diverges from the Federal Aviation Administration‑style framework that Anthropic’s CEO outlined earlier this summer. Among the contentious elements is a provision for an “AI kill switch,” a mechanism that safety advocates argue is insufficiently defined and could be easily circumvented. The draft also sets a lower threshold for triggering federal oversight, prompting concerns that advanced systems could escape rigorous review.
The controversy matters because federal AI policy is coalescing around a few core principles—risk‑based licensing, transparency, and robust mitigation tools. If Anthropic’s softer version gains legislative foothold, it could set a precedent for industry‑driven standards that fall short of the broader security objectives embodied in the FRONTIER Act. Lawmakers and regulators are now faced with reconciling competing drafts while maintaining bipartisan momentum on AI governance.
Stakeholders will be watching for an official response from Anthropic, potential amendments to the leaked text, and how the debate shapes upcoming hearings on the FRONTIER Act. The next few weeks could determine whether industry‑sponsored proposals will complement or undercut the emerging federal framework for advanced AI safety.
Sources
Back to AIPULSEN