Anthropic · Anthropic API Usage Policy · View original document ↗

Weapons of Mass Destruction Prohibition

High severity High confidence Explicitdocumentlanguage Rare · 1 of 325 platforms
Share 𝕏 Share in Share 🔒 PDF
Monitor governance changes for Anthropic Create a free account to receive the weekly governance digest and monitor one platform for governance changes.
Create free account No credit card required.
Document Record

What it is

Claude cannot be used to help create, design, or acquire weapons of mass destruction including biological, chemical, radiological, or nuclear weapons, or any modifications to evade detection.

This analysis describes what Anthropic's agreement states, permits, or reserves. It does not constitute a legal determination about enforceability. Regulatory applicability and practical outcomes may vary by jurisdiction, enforcement context, and individual circumstances. Read our methodology

ConductAtlas Analysis

Why it matters (compliance & governance perspective)

This is one of the absolute prohibitions in the policy, covering not just direct weapon design but also precursor development, weaponization processes, and evasion modifications, making the scope of the prohibition comprehensive.

Recent Activity

This document changed recently

High Feb 27, 2026

Defense contractors and federal agencies using Claude must find alternatives. Enterprise customers with defense-adjacent business face compliance risk.

Consumer impact (what this means for users)

This provision means Claude is hard-blocked from assisting with any aspect of weapons of mass destruction development. For the vast majority of users, this has no practical impact, but it is a categorical prohibition that cannot be unlocked through any operator permission.

How other platforms handle this

Runway Medium

You may not use Runway's tools to create content that promotes, glorifies, or facilitates acts of terrorism, mass violence, or genocide, or that could be used to provide material support to individuals or organizations engaged in such activities.

Mistral AI Medium

Customer will not, and will not permit any other person (including any End User) to: ... (d) attempt to reverse engineer, decompile, or otherwise attempt to discover the source code or underlying components (e.g., algorithms, weights, or systems) of the Mistral AI Products, including using the Outpu...

Perplexity AI Medium

You may not use the Services to attempt to circumvent, disable, or otherwise interfere with safety-related features of the Services, including features that prevent or restrict the generation of certain types of content.

See all platforms with this clause type →

Monitoring

Anthropic has changed this document before.

Receive same-day alerts, structured change summaries, and monitoring for up to 10 platforms.

Start Watcher free trial Or create a free account →
▸ View Original Clause Language DOCUMENT RECORD
"
Synthesize, or otherwise develop, high-yield explosives or biological, chemical, radiological, or nuclear weapons or their precursors, including modifications to evade detection or medical countermeasures [...] Produce, modify, design, or illegally acquire weapons, explosives, dangerous materials or other systems designed to cause harm to or loss of human life.

— Excerpt from Anthropic's Anthropic API Usage Policy

ConductAtlas Analysis

Institutional analysis (Compliance & governance intelligence)

(1) REGULATORY LANDSCAPE: This provision engages the Biological Weapons Anti-Terrorism Act, Chemical Weapons Convention implementation statute, Nuclear Non-Proliferation frameworks, and export control regulations including the Export Administration Regulations (EAR) and International Traffic in Arms Regulations (ITAR). The Department of Justice, Department of State, and Department of Commerce have overlapping enforcement authority. Internationally, UN Security Council Resolution 1540 obligations apply. (2) GOVERNANCE EXPOSURE: High reputational and legal risk if violations occur; low operational compliance burden for standard users. The inclusion of 'precursors' and 'modifications to evade detection or medical countermeasures' reflects awareness of dual-use research risks. For research institutions and biosecurity organizations, the policy's application to legitimate defensive research may require specific guidance from Anthropic. (3) JURISDICTION FLAGS: Export control implications apply globally. US persons and entities have specific obligations under EAR and ITAR regardless of where they are located. Academic and research institutions may face tension between this prohibition and legitimate biosecurity research, which is jurisdiction-dependent in its legal treatment. (4) CONTRACTUAL AND VENDOR IMPLICATIONS: Research institutions, defense contractors, and biosecurity organizations using Claude must evaluate whether their intended use cases fall within the prohibition or legitimate research exceptions. The governmental carve-out may be relevant for certain defense or public health research applications, but the CBRN weapons prohibition appears to be among the absolute limits not subject to modification. (5) COMPLIANCE CONSIDERATIONS: Organizations in dual-use research fields should obtain specific written guidance from Anthropic on permitted biosecurity research use cases before deployment. Legal teams should assess whether existing export control compliance programs extend to AI tool usage in relevant research contexts.

Full compliance analysis

Regulatory citations, enforcement risk, and due diligence action items.

Track 1 platform — free Try Watcher free for 14 days

Free: track 1 platform + weekly digest. Watcher: 10 platforms + same-day alerts. No credit card required.

Applicable agencies

  • FTC
    The FTC has general consumer protection authority, though primary enforcement for WMD-related violations would rest with DOJ and national security agencies not listed in the available agency options
    File a complaint →

Applicable regulations

CFAA
United States Federal
DMCA
United States Federal
DSA
European Union
Trump Executive Order on AI Policy Framework
US

Provision details

Document information
Document
Anthropic API Usage Policy
Entity
Anthropic
Document last updated
May 11, 2026
Tracking information
First tracked
May 11, 2026
Last verified
May 11, 2026
Record ID
CA-P-009968
Document ID
CA-D-00013
Evidence Provenance
Source URL
Wayback Machine
Content hash (SHA-256)
60e693438d9f7f47deb8f3bfb819343e26b5fe0eb90d56280568f1dd95ae660f
Analysis generated
May 11, 2026 00:39 UTC
Methodology
Evidence
✓ Snapshot stored   ✓ Hash verified
Citation Record
Entity: Anthropic
Document: Anthropic API Usage Policy
Record ID: CA-P-009968
Captured: 2026-05-11 00:39:26 UTC
SHA-256: 60e693438d9f7f47…
URL: https://conductatlas.com/platform/anthropic/anthropic-api-usage-policy/weapons-of-mass-destruction-prohibition/
Accessed: May 13, 2026
Permanent archival reference. Stable identifier suitable for legal filings, compliance documentation, and research citation.
Classification
Severity
High
Categories

Other risks in this policy

Professional Governance Intelligence

Need to monitor specific governance provisions?

Professional includes provision-level monitoring, governance timelines, regulatory mapping, and audit-ready analysis.

Arbitration clauses AI governance Data rights Indemnification Retention policies
Start Professional free trial

Or start with Watcher →

Built from archived source documents, structured governance mappings, and historical version tracking.

Frequently Asked Questions

What does Anthropic's Weapons of Mass Destruction Prohibition clause do?

This is one of the absolute prohibitions in the policy, covering not just direct weapon design but also precursor development, weaponization processes, and evasion modifications, making the scope of the prohibition comprehensive.

How does this clause affect you?

This provision means Claude is hard-blocked from assisting with any aspect of weapons of mass destruction development. For the vast majority of users, this has no practical impact, but it is a categorical prohibition that cannot be unlocked through any operator permission.

How many platforms have this type of clause?

ConductAtlas has identified this type of provision across 1 platforms. See the full comparison.

Is ConductAtlas affiliated with Anthropic?

No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by Anthropic.