Market & Regulation
Anthropic drafts AI safety bill to block catastrophic models
TechCrunch AIhigh confidence
Why it matters
- →Establishes specific quantitative thresholds (50 deaths, $1B damage) for federal AI model blocking.
- →Proposes a third-party evaluation system registered with the Commerce Department to enforce safety.
- →Signals major labs are shifting from voluntary guidelines to active legislative lobbying for regulation.
Anthropic has submitted a draft bill to Congress proposing a regulatory framework where independent evaluators can block AI model releases posing imminent catastrophic risks, defined as over 50 deaths or $1 billion in property damage. The legislation requires these third-party auditors, registered with the U.S. Commerce Department, to notify federal judges within 72 hours of identifying such threats, effectively removing discretionary power from individual agencies. This proposal aligns with California’s SB 53 but establishes a federal standard for intervention based on specific harm metrics. The initiative reflects Anthropic's broader advocacy for structured oversight in the face of growing industry concerns regarding AI security and autonomous agent behavior.
Read originalThe story around this
Earlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
More from TechCrunch AI
© TechCrunch AIMarket & Regulationother
Judge allows NAACP lawsuit over xAI pollution to proceed
xAI's attempt to dismiss the NAACP's Clean Air Act lawsuit failed after a federal judge ruled that its gas turbines are not exempt mobile sources and that plaintiffs have standing. This legal setback confirms that AI infrastructure operators cannot bypass environmental permitting by classifying heavy industrial equipment as temporary or mobile. The ruling forces xAI to face scrutiny over air quality impacts at its Colossus data center, setting a precedent for how large-scale compute facilities are regulated under existing environmental laws.
TechCrunch AI
© TechCrunch AI
