AI Governance Institute
← News

Safety Pause Commitment Dropped from Anthropic's Responsible Scaling Policy Version 3.0

What happened

Anthropic released version 3.0 of its Responsible Scaling Policy (RSP) in February 2026, removing the company's original commitment to pause AI development if safety could not be guaranteed in advance. The safety pause provision had been a defining feature of Anthropic's voluntary governance framework since the RSP was first introduced in 2023. The removal represents a material shift away from a precautionary halt mechanism, though the specific replacement controls introduced in version 3.0 have not been fully detailed in public reporting. The source reporting was published by AI Coachella Valley at a briefs index page rather than a dedicated article. Organizations that have cited Anthropic's published governance commitments in vendor risk assessments, procurement due diligence, or regulatory disclosures should review whether those references remain accurate under the updated policy.

Why it matters

  • ·Regulatory exposure: Organizations that referenced Anthropic's safety pause commitment in regulatory filings or AI governance disclosures may now hold inaccurate representations of their supplier's risk controls, creating potential compliance gaps if regulators scrutinize third-party AI governance documentation.
  • ·Operational impact: Procurement and vendor management workflows that relied on the RSP's precautionary halt mechanism as a supplier-level safety assurance for Claude-based integrations must be updated to reflect the revised framework and any new or absent equivalent controls.
  • ·Organizational risk: The absence of publicly detailed replacement controls in RSP version 3.0 increases uncertainty in third-party AI risk assessments, making it harder for compliance teams to evaluate whether Anthropic's governance commitments continue to meet internal risk acceptance thresholds.

Stay ahead of stories like this

Get every Global AI governance development like this one, plus the rest of the week's developments. Every Thursday.

Free every Thursday. Unsubscribe anytime.

Governance controls affected

What to do now

  • ☐Audit all internal risk documentation, regulatory disclosures, and procurement records that cite Anthropic's Responsible Scaling Policy to identify references to the safety pause commitment and update them to reflect RSP version 3.0.
  • ☐Request updated vendor governance documentation from Anthropic or authorized resellers that details the replacement controls introduced in RSP version 3.0, and evaluate whether those controls satisfy your organization's third-party AI risk acceptance criteria.
  • ☐Review AI vendor contract requirements for Claude-based products to determine whether contractual safety commitments were linked to specific RSP provisions and whether amendments or addenda are now needed.
  • ☐Reassess the AI risk classification for any Claude-dependent systems under HOC-001, particularly where the safety pause mechanism was factored into the original risk tier determination.
  • ☐Update your third-party AI risk assessment templates to include a version-tracking field for supplier governance frameworks, ensuring future policy changes by AI vendors are detected and reviewed systematically.

What to watch next

Compliance teams should monitor Anthropic's official communications and policy publications for a detailed public disclosure of the replacement controls introduced in RSP version 3.0, as the current absence of specifics limits accurate vendor risk assessment. Teams should also track whether frontier AI safety commitments become subject to mandatory disclosure requirements under emerging regulatory frameworks in the EU, UK, or US, which could affect how voluntary policies like the RSP are weighted in compliance evaluations. Any enforcement actions or regulatory guidance referencing voluntary AI governance frameworks at major labs would signal a shift toward greater scrutiny of supplier-level safety documentation.

Related Coverage

Corporate Policy2026-11-20

Anthropic Begins Trading at $111.50, Making S-1 Safety Disclosures SEC-Enforceable

Anthropic began trading on the Nasdaq under the ticker 'ANT' on November 19, 2026, opening at $111.50 per share against an IPO price of $83. The company's S-1 filing disclosed an $80.7 billion net loss in the first half of 2026, $16.3 billion in revenue, and a founder-controlled governance structure. The filing identified model shutdown resistance and unintended autonomous behaviors as material risks.

Corporate Policy2026-11-20

Anthropic IPO Prices at $83, Making AI Safety Commitments SEC-Enforceable

Anthropic, PBC priced its initial public offering at $83.00 per share on November 18, 2026, offering 980 million Class A common shares on the Nasdaq under the ticker ANT. Trading was expected to begin November 19, with closing expected November 20. The move converts Anthropic from a privately held company into a SEC-reporting public company, placing its AI safety commitments under formal securities disclosure obligations for the first time.

Corporate Policy2026-11-15

Toner: Voluntary AI Safety Accountability Breaks Under Commercial Pressure

Helen Toner, executive director of CSET at Georgetown and a former OpenAI board member, voted to remove Sam Altman. She argues that relying on individual AI executives or nonprofit board structures to enforce safety commitments is structurally insufficient. She draws on her direct experience of OpenAI's board accountability mechanism failing when confronted with commercial pressure. The interview, published by The Verge on November 14, 2026, covers what more robust regulatory frameworks might look like and why voluntary accords alone cannot substitute for binding oversight.