An AI safety company that develops and operates large language models, including the Claude family of AI assistants used by consumers and businesses for text generation, analysis, and conversation. The company's policies govern how user data is collected and processed during AI interactions, as well as what types of content and use cases are prohibited on their platforms. These policies are significant for users because they determine privacy protections for sensitive conversations with AI systems and establish boundaries for acceptable AI-assisted activities.

11 Documents
143 All provisions 29 High severity 90 Medium severity 24 Low severity
6 Changes · 12 mo 1 high
Jul 14, 2026 Last change

High, provisions that significantly limit your legal rights, authorize broad data collection, or create material financial exposure. Medium, provisions worth knowing about but with partial protections or limited scope. Low, standard terms with minimal consumer impact.

Governance Coverage
Acceptable Use
111 provisions · 91 high · 20 med
Contract Terms
61 provisions · 23 high · 26 med
Data Sharing
51 provisions · 13 high · 33 med
Privacy Rights
48 provisions · 3 high · 27 med
Liability Limitation
43 provisions · 20 high · 19 med
Account Control
37 provisions · 23 high · 11 med
Payment Fees
34 provisions · 16 high · 18 med
Data Usage
31 provisions · 11 high · 13 med
Data Collection
30 provisions · 6 high · 16 med
Refunds
28 provisions · 15 high · 10 med
Quick Facts
Arbitration: Present
AI Training: Disclosed (high impact)
Data Retention: Specified
Key Governance Provisions
High Exclusion of All Consequential and Punitive Damages

The clause purports to eliminate Anthropic's and its affiliates' financial liability across virtually every recognized category of damages, significantly limiting users' legal recourse.

High Risk Reports Quantify Risk Across All Deployed Models

A requirement to quantify risk across all deployed models — rather than selected ones — establishes a comprehensive scope for Anthropic's internal risk accounting obligations.

High Multi-Party Authorization for Production Model Weight Access

The requirement directly restricts who can access model weights and under what conditions, establishing a structural safeguard against unauthorized or unilateral access to core AI assets.

High ASL-3 Deployment Safeguard CBRN Misuse Prevention

The clause identifies CBRN misuse as the primary threat model driving ASL-3 deployment safeguard design, establishing the scope of harm Anthropic is specifically targeting at that capability level.

High Misalignment Risk Affirmative Case Requirement at AI R&D-4

The requirement creates a formal, documented obligation for Anthropic to affirmatively demonstrate misalignment risk awareness and mitigation at a specific capability level, rather than leaving it to…

Document Coverage
Acceptableusepolicy Enterpriseterms Privacypolicy Responsibleai Subprocessorlist Systemcard Termsofservice
Last change detected: Jul 14, 2026 Changes (12 months): 6 (1 high) Documents monitored: 11 Provisions tracked: 143
Compare With Similar Platforms
AI21 Labs Anyscale Apple Intelligence Baseten Cerebras Cohere Copy.ai Cursor

Documents

Anthropic Sub-Processors

Low

This document lists the outside companies Anthropic sends user data to, what each company does with it, and which Anthropic products are involved. Every Anthropic product routes data through at …

Last updated July 6, 2026 · Sub processor list
10 provisions 3 versions captured Version history →

Claude Sonnet 5 System Card

Medium

This is Anthropic's system card for Claude Sonnet 4.5, a technical disclosure document describing how the model is built, evaluated, and constrained. The document states that certain behaviors are hardcoded …

Last updated July 6, 2026 · System card
7 provisions 4 versions captured Version history →

Claude Opus 4.8 System Card

High

This is Anthropic's system card for Claude Opus 4.8, a large language model released in May 2026, documenting pre-deployment safety evaluations across chemical and biological weapons risks, cybersecurity, agentic safety, …

Last updated July 6, 2026 · System card
8 provisions 2 versions captured Version history →

Anthropic Responsible Scaling Policy

Medium

This document is Anthropic's internal policy for deciding when its AI systems are safe enough to keep developing and deploying. It sets specific capability thresholds that trigger safety requirements — …

Last updated May 12, 2026 · Responsible ai
10 provisions 1 version captured Version history →

Anthropic Consumer Terms

Medium

These are the rules governing your use of Anthropic's consumer Services. Anthropic can change, suspend, or cut off your access at any time without warning, and its financial liability to …

Last updated May 12, 2026 · Terms of service
8 provisions 1 version captured Version history →

Anthropic API Usage Policy

Medium

This policy sets the rules for what users and operators can and cannot do with Anthropic's products and services. If you use Anthropic's tools to interact with or advise consumers, …

Last updated May 11, 2026 · Acceptable use policy
9 provisions 2 versions captured 1 significant change Version history →

Anthropic Commercial Terms

Medium

This document establishes the commercial terms governing business and developer access to Anthropic's Claude API, excluding individual Claude.ai users who are covered under separate terms. The agreement specifies that customers …

Last updated May 11, 2026 · Enterprise terms
10 provisions 1 version captured Version history →

Anthropic Privacy Policy

Medium

This policy explains how Anthropic collects and uses your information, including using your conversations to train its AI by default unless you opt out in your account settings. Anthropic transfers …

Last updated May 5, 2026 · Privacy policy
8 provisions 3 versions captured Version history →

Anthropic API Terms

Medium

These terms govern your use of Anthropic's API and place most of the risk on you: Anthropic can shut down or change the service at any time without notice, and …

Last updated May 5, 2026 · Terms of service
10 provisions 1 version captured Version history →

Claude.ai Terms of Service

Medium

This document sets the rules for using Claude.ai: Anthropic can change or shut down the service at any time without warning, and your content may be used to train AI …

Last updated May 5, 2026 · Terms of service
10 provisions 2 versions captured Version history →

Anthropic Privacy Policy (Superseded Capture)

Medium

This document explains how Anthropic collects and uses your data, including the content of your conversations with Claude. By default, Anthropic uses your chats to train its AI—you can turn …

Last updated May 5, 2026 · Privacy policy
9 provisions 4 versions captured Version history →

Recent Changes

February 27, 2026 · Anthropic API Usage Policy High

The Pentagon has designated Anthropic as a supply chain risk to national security under 10 USC 3252 after the company refused to remove acceptable use policy restrictions that prohibit mass …

View change record →
📈 Policy Drift Analysis View all changes →

Medium Severity Provisions (69)

View all high severity provisions → Compare Anthropic with other platforms →

Applicable Regulations

EU AI Act
European Union
View official text ↗
BIPA
Illinois, USA
View official text ↗
California AB 2013 AI Training Data Transparency
US-CA
View official text ↗
CCPA/CPRA
California, USA
View official text ↗
Colorado AI Act
US-CO
View official text ↗
CFAA
United States Federal
View official text ↗
Connecticut Data Privacy Act Amendments
US-CT
View official text ↗
CAN-SPAM
United States Federal
View official text ↗
DMCA
United States Federal
View official text ↗
DSA
European Union
View official text ↗
ePrivacy Directive
European Union
View official text ↗
EU AI Act - High Risk Provisions
EU
View official text ↗
FTC Act Section 5
United States Federal
View official text ↗
GDPR
European Union
View official text ↗
Indiana Consumer Data Protection Act
US-IN
View official text ↗
Kentucky Consumer Data Protection Act
US-KY
View official text ↗
Texas AI Act
Texas, USA
View official text ↗
Trump Executive Order on AI Policy Framework
US
View official text ↗
UK GDPR
United Kingdom
View official text ↗
Universal Opt-Out Mechanism Expansion 2026
US
View official text ↗

Related Analysis

Privacy · April 14, 2026
Deleted Claude Conversations Aren't Gone for 30 Days

Anthropic is more transparent than most AI companies about data retention. Here's exactly what happens when you delete your data, and how t…

Compare with Similar Platforms

AI21 Labs
ai
Anyscale
ai
Apple Intelligence
ai
Baseten
ai
Cerebras
ai
Cohere
ai
Copy.ai
ai
Cursor
ai

Frequently Asked Questions

What Anthropic documents does ConductAtlas track?

ConductAtlas tracks 11 Anthropic documents including terms of service, privacy policy, and other governance documents. Every document is captured daily with cryptographic verification.

How many policy changes has Anthropic made in the past year?

Anthropic has made 6 policy changes in the past 12 months across the documents ConductAtlas tracks, including 1 classified as high severity.

How many provisions has ConductAtlas classified for Anthropic?

ConductAtlas has classified 143 provisions across Anthropic's tracked documents. 29 are rated high severity, 90 medium, and 24 low.

Can I get alerts when Anthropic changes their policies?

Yes. Monitor subscribers ($4.99/month) can add Anthropic to their watchlist and receive same-day email alerts whenever any tracked document changes.