Thursday, 22 October 2026
Anawords
Research desk
Research noteResearched 22 Oct 2026 · published 22 Oct 2026, 17:39 IST

Research note: Anthropic circulates draft US AI safety bill that would let a court halt a model posing imminent catastrophic risk

How we checked the details of Anthropic’s draft bill and kept views on it balanced.

Research led by Mayank Sahu
Compiled with AI assistance (Claude by Anthropic)
Read the story →

At a glance

1searches run
3sources read
2sources used
6facts labelled
Strongevidence strength · 4 of 5 facts (80%)

How we rate evidence: Most facts come from official sources or two or more independent outlets. We count the share of known facts labelled Confirmed or Reported: 70% or more is Strong, 40–69% Mixed, below 40% Limited.

From sources to story

How many sources we read, how many we used, and how many are official or archived.

Read3Every source we opened
Used2Relied on for the story
Cited2Linked in the story
Official0From the company or a public body
Archived2Saved on the Internet Archive

How sure we are, fact by fact

  • Reported · 4 — Two or more independent, named outlets say it.
  • Unconfirmed · 1 — One report only, shown with its source.
  • Not known · 1 — Not announced or not answered yet.

Evidence map

Each row is a fact from the story; each column is a source. A filled dot means that source supports the fact.

FactLabel1OUT2OUT
Anthropic has circulated a draft federal AI safety bill; Politico first reported it on 16 October 2026.Reported
The threshold is a foreseeable, imminent risk of more than 50 deaths or $1 billion in damage.Reported
Evaluators registered with Commerce must report an imminent risk within 72 hours.Reported
Commerce would ask a federal judge in Washington, DC, to halt the model.Reported
Some state AI laws would be preempted.Unconfirmed
Whether any lawmaker will introduce the bill.Not known
Question

What we set out to check

What the draft says, who would decide, and what questions it raises.

Findings

What we found

  • Both outlets agree on the thresholds, the three scenarios, the 72-hour notice, the Commerce-only role and the DC court.
  • Gadget Review adds transparency, whistleblower and preemption provisions.
  • Neither outlet reports reactions from lawmakers.
Discrepancies

Numbers that differ

None.

Method

Labels we chose and why

  • Reported for the core mechanism, in both outlets.
  • Unconfirmed for preemption and the extra provisions, from one outlet.
  • We offer both supportive and critical framings, without taking a side.
Open questions

What we could not verify

  • The draft text and Politico’s original report.

Can you help confirm one of these? Send us a source. When we find an answer, we update this note and the story, with the date.

Next steps

How we could improve this

Obtain the draft or Anthropic’s framework document and confirm the provisions directly.

Research timeline

  1. Research22 Oct 2026
  2. Research note published22 Oct 2026, 17:39 IST
  3. Story published22 Oct 2026, 17:49 IST

What we searched

  • Anthropic AI safety bill 50 deaths $1 billion Commerce Department federal judge

Every source we read

SourceUsed?Why
TechCrunch
Archived copy		USEDThresholds, process, questions, Anthropic’s non-response.
Gadget Review
Archived copy		USEDAdditional provisions and preemption.
Technology.orgNOT USEDCould not open.

Sources cited in the story

  1. 1TechCrunch: Anthropic pitches AI safety bill that may block a model if 50+ people could be killedOUTLET
  2. 2Gadget Review: Anthropic’s proposed AI safety bill only kicks in at 50 deaths or $1 billion in damagesOUTLET

Research notes show our working. They are not peer-reviewed papers. Data: facts as JSON · How we report