Anthropic's Leaked AI Bill Draws Backlash

Episode 1 · Anthropic's secretly pushed AI safety bill draws fire from White House and safety researchers (2026-10-17, 29 posts)

Politico's Owen Dahlkamp exclusively reported that Anthropic has been privately pitching U.S. lawmakers an AI safety bill it helped draft, with the draft text leaking online afterward. The rare move of a frontier lab writing its own regulatory text drew fierce criticism from the White House, safety researchers, and developers, becoming the week's biggest AI policy controversy. Neither the bill's final terms nor its legislative path are settled, and markets price its passage odds very low.

Confirmed

• Per Politico, the bill closely mirrors Anthropic's June framework. Under its terms, the government could only block a model's release if it posed foreseeable, imminent risk of over $1 billion in losses or 50+ deaths/serious injuries, with risk limited to bioweapons, loss of control, or cyberattack—wording similar to California's SB 53. Compared with the FRONTIER Act, it drops licensing and a dedicated AI regulator while keeping framework + transparency reports + incident reporting + third-party assessments.
• White House AI chief David Sacks attacked the bill's 10^25 FLOPs definition of frontier models as an outdated 2023 EU standard that would sweep in much of Silicon Valley; he amplified developer sagivo's point that the cumulative-compute definition would burden fine-tuners, since post-training is now the industry's main activity.
• Safety researchers reacted angrily: Nathan Calvin called the draft physically enraging, arguing its preemption would crush SB 53-style state protections and leave the U.S. less protected against catastrophic risk, with auditors requiring no government licensing. Jay Shooster called it outrageous and unrepresentative of the AI safety community.
• Gary Marcus amplified @So8res's critique that courts would find it harder to stop dangerous AI research, and accused the bill of securing substantial liability shields—e.g., immunity despite hundreds of millions in losses or dozens of deaths. François Fleuret and others mocked that Anthropic could theoretically 'kill 49 people and steal up to $1 billion' without legal consequence.
• Anthropic's own safety researcher Sneha Revanur criticized the plan to tell Congress that blocking dangerous deployments should be harder than today. Kelsey Tuoc argued Anthropic truly fears flexible executive-branch shutdown power; Zvi called the leak unacceptable and disturbing over preemption and weak enforcement; NY Assemblymember Alex Bores also raised open-source concerns.
• Polymarket prices the chance of a U.S. AI safety bill passing by end-2026 at only 6%-9%.

Unconfirmed

• The bill's final terms and legislative path remain uncertain; low Polymarket pricing suggests markets doubt near-term passage.

Why it matters

• This is the first exposed case of a frontier lab directly drafting federal legislation for its own industry, raising fears of regulatory capture—so much so that Anthropic's own safety staff dissented publicly. Sacks's rebuke signals open White-House–Anthropic rifts, and the safety community's split foreshadows an intensified federal AI-legislation fight.

• Anthropic wrote its own frontier AI regulation bill and is pitching it to Congress, Politico reports — Promptmethus · 2026-10-17
• Anthropic quietly pitches its own AI safety bill — and it guts agency power — ShakeelHashim · 2026-10-17
• AI safety figure slams new bill as not representing the safety community — ShakeelHashim · 2026-10-17
• Anthropic pitched its own AI safety bill to Congress; Polymarket odds say 6% — Polymarket · 2026-10-17
• SCOOP: Anthropic quietly pitching its own AI safety bill to Congress — asusarla · 2026-10-17
• Anthropic quietly pitches its own AI safety bill; ex-safety lead calls it a step back — Miles_Brundage · 2026-10-17
• Kelsey Tuoc: Anthropic's draft bill reveals fear of executive power, not safety — Miles_Brundage · 2026-10-17
• AI safety researchers slam Anthropic-backed bill for lacking auditor licensing — Miles_Brundage · 2026-10-17
• AI safety researchers slam Anthropic-backed federal AI bill as 'astonishingly bad' — michael_nielsen · 2026-10-17
• Draft Anthropic-backed AI regulation bill text now circulating online — ShakeelHashim · 2026-10-17
• Anthropic quietly pitches its own AI safety bill to lawmakers, drawing fire over pace — sjgadler · 2026-10-17
• Anthropic quietly pitches its own AI safety bill to lawmakers, Politico reports — AlexTensor · 2026-10-17
• Anthropic's AI bill: independent audits, 72-hour risk reports, judge-ordered model suspensions — AlexTensor · 2026-10-17
• Zvi: Leaked Anthropic-Backed AI Regulatory Bill Is Unacceptable, Weakens Enforcement — TheZvi · 2026-10-17
• David Sacks slams Anthropic-backed safety bill: 10^25 FLOPs threshold would sweep in most of Silicon Valley — DavidSacks · 2026-10-18
• Critics slam Anthropic for backing bill that lets AI labs define their own safety thresholds — HankYeomans · 2026-10-18
• Sacks warns Anthropic-drafted safety bill's 10^25 FLOPs frontier threshold would sweep in most of Silicon Valley — DavidSacks · 2026-10-18
• Developer warns cumulative-FLOPs threshold would force weekly fine-tuners into compliance paperwork — DavidSacks · 2026-10-18
• Satire hits Anthropic's proposed AI safety bill: 'labs writing their own rules' — james_mtc · 2026-10-18
• Sacks slams Anthropic-linked safety bill's 10²⁵ FLOPs frontier threshold; critics say cap capabilities, not FLOPs — jeremyakahn · 2026-10-18

9 more related posts →

Episode 2 · Anthropic lobbies Congress to bar AI intervention unless losses top $1B or 50 casualties (2026-10-19, 2 posts)

Anthropic is privately lobbying the US Congress on a bill that would let the government block a "superintelligence model" only if it could cause over $1 billion in losses or kill/injure more than 50 people. The proposal has sparked debate over whether it effectively shields AI firms from oversight.

• Anthropic privately pitching Congress a bill: government can block an SI model only above $1B damage or 50 deaths — AIFlow_ML · 2026-10-19
• Anthropic proposes blocking AI models only over $1B damage or 50+ casualties — fsharpman · 2026-10-19

Episode 3 · Anthropic's Leaked Federal AI Bill Draft Slammed as Regulatory Capture (2026-10-20, 3 posts)

Anthropic's leaked federal AI bill draft, reported by Politico, was criticized as far weaker than existing proposals, with law professor Stephen Casper calling it textbook regulatory capture. OpenAI policy chief Chris Lehane meanwhile urged Congress to act on AI legislation.

• OpenAI's Chris Lehane pushes Congress to act as Anthropic leak roils AI safety world — ShakeelHashim · 2026-10-20
• Leaked Anthropic draft AI bill panned as weaker than FRONTIER Act, 'safety-first' pitch rings hollow — ShakeelHashim · 2026-10-20
• Law professor slams leaked Anthropic draft AI bill as textbook regulatory capture — StephenLCasper · 2026-10-20