US vice-president says AI developers should not create risks then ask for rules

single source· 1 articles · confidence: low · first seen 2026-09-16 10:42 UTC

What this means for you

Nothing to act on — this is a speech, not a rule: no bill, no agency guidance, no date. The one thing worth recording is that the US executive branch's stated position is that developers, not government, should manage the risks of their most capable models.

JD Vance, the US vice-president, has rejected calls for government regulation of AI safety, telling companies that build the most advanced models that if they are "building Frankenstein" they should stop. The remarks were addressed to Anthropic co-founder Dario Amodei, who has asked Washington to coordinate controls on AI systems, including with China. The Guardian reported them on 16 September. The same piece notes a former Anthropic researcher revisiting a recent claim that AI could destroy humanity. No bill, rule or date accompanied the comments.

Key facts

  • ·US vice-president JD Vance said companies building advanced AI models should stop rather than come to government asking for regulation. source
  • ·Vance's remarks were addressed to Anthropic co-founder Dario Amodei, who has called on Washington to coordinate control of AI systems, including with China. source
  • ·The Guardian published the report on 16 September 2026. source
  • ·The report also notes a former Anthropic researcher revisiting a recent claim that AI could destroy humanity. source
  • ·No legislation, regulation or policy change was announced alongside the comments. source

What the sources say

  • The Guardian AI — Short report on the vice-president's rebuke of Anthropic's co-founder and the safety-regulation camp.

Sources

The original reporting. Follow these — they did the work.

← the wire