1. What the document is: Anthropic's "Usage Policy" (also called the Acceptable Use Policy) — Universal Usage Standards (general conduct rules), High-Risk Use Case Requirements (a human must review AI output, and its use must be disclosed, in legal/healthcare/finance/employment/insurance/admissions/journalism contexts), Additional Use Case Guidelines (chatbots, minors, agentic use, MCP servers), and a Minors-Serving Organizations addendum.

What this wiki found — complete, every page

AI trainingRoyalty-freeSublicensablePerpetual / irrevocableTransferableDerivative works

Presence facts, not verdicts — each flagged term links to its definition and the exact clause on this page. Absence of a badge means the term isn't currently flagged here, not that the page is risk-free.

Claude.ai – Acceptable Use Policy

Dataset: GenGA (Generative AI Governance Archive) — 11 AI services, 2025–present

GenGA (Generative AI Governance Archive) is this wiki’s Generative-AI dataset: raw policy snapshots captured by the Open Terms Archive genai-eu project across 11 GenAI providers (ChatGPT, Claude.ai, DeepSeek, Google Generative AI Services, Le Chat, Llama API, Meta AI, Microsoft Copilot, Perplexity, Qwen Chat, xAI). Unlike PGAv2, GenGA has no pre-tagged risk scores — all risk scoring and keyword tagging on these pages is LLM-assigned by direct reading, spanning 2025–present.

Note: This page contains documented policy clauses. Risk assessment is qualitative and context-dependent. For analysis of patterns across platforms, see:

Source: sources/GenGA/Claude.ai/Acceptable Use Policy/ (raw Markdown captures, Open Terms Archive genai-eu dataset — no pre-tagged JSONL exists for this source) Last updated: 2025-11-11 (only capture in this dataset)

Note on methodology: No pre-computed risk_score/keyword fields exist for this source. Keyword tags and clause analysis on this page are LLM-assigned, using the same rubric/vocabulary as the PGAv2 pages plus new GenAI-specific tags (competing model ban, ai disclosure). Treat as first-pass analysis, not externally verified ground truth. This wiki’s risk-scoring system was retired project-wide on 2026-06-21 (see methodology.md §4) — no page, GenGA included, computes or displays a numeric risk score.

Overview

1. What the document is: Anthropic’s “Usage Policy” (also called the Acceptable Use Policy) — Universal Usage Standards (general conduct rules), High-Risk Use Case Requirements (a human must review AI output, and its use must be disclosed, in legal/healthcare/finance/employment/insurance/admissions/journalism contexts), Additional Use Case Guidelines (chatbots, minors, agentic use, MCP servers), and a Minors-Serving Organizations addendum.

2. Input rights: Not addressed — this is a conduct policy, not a content-licensing document.

3. Output restrictions: The “Do Not Abuse our Platform” section bans “Utilization of inputs and outputs to train an AI model (e.g., ‘model scraping’ or ‘model distillation’) without prior authorization from Anthropic.” That’s a direct, explicitly-named ban on model distillation — a technique for training a smaller model to copy Claude’s behavior — named more precisely here than most competing-model bans elsewhere in this wiki. Also bans jailbreaking or prompt injection without Anthropic’s authorization.

4. Non-explicit predatory clauses: None found — Anthropic’s AUP doesn’t use “improve services” framing or claim rights to user Feedback; it’s purely about user conduct.

5. Regulatory references: COPPA (the US Children’s Online Privacy Protection Act) is named explicitly in the minors-serving-organizations addendum. No EU AI Act or GDPR reference found.

6. Regional variation: Not stated explicitly in this document (no “applies if you reside in…” language), though the page shows an “English” language selector, suggesting localized versions may exist outside this capture.

7. Key risk to users: Low-to-moderate. The explicit model-distillation ban is the most consequential clause for developers and researchers — it rules out a common technique for building smaller models from Claude’s outputs. The mandatory AI-disclosure requirements for consumer chatbots and high-risk use cases actually protect users rather than harm them; they’re flagged here because they’re a recurring transparency pattern worth comparing across providers.

Flagged Keywords & Risks (LLM-assigned)

competing model ban — “Utilization of inputs and outputs to train an AI model (e.g., ‘model scraping’ or ‘model distillation’) without prior authorization from Anthropic.” Why it matters: one of the most explicitly-named anti-distillation clauses in this dataset — a useful comparison point against OpenAI’s broader, less technical “develop models that compete with OpenAI” language.
ai disclosure — mandatory AI-disclosure for all consumer-facing chatbots (“must disclose to users that they are interacting with AI rather than a human… at a minimum at the beginning of each chat session”) and for High-Risk Use Case outputs shown to individuals or consumers. Why it matters: one of the most concrete, mandatory (not just recommended) AI-disclosure requirements found in this dataset so far.

Regulatory & Research Context

Pandit et al. (2026), whose six-provider analysis includes Anthropic’s Claude by name, would likely read the explicit “model scraping”/“model distillation” ban here as a sharper, more technically specific instance of the kind of output-use restriction their unfairness framework examines, compared to the broader “competing model” language found in other providers’ terms. The mandatory AI-disclosure requirement for consumer-facing chatbots is consistent with Pandit et al. (2026)‘s observation that even where terms shift responsibility onto users or third-party deployers, transparency-style obligations can coexist with that shifted risk; Edwards et al. (2025) would frame the High-Risk Use Case human-in-the-loop requirements as part of the “platformisation paradigm,” in which Anthropic defines the governance terms under which its own model may be used for consequential decisions.

Changes Summary

Date	What changed
2025-11-11	Baseline and only version captured in this dataset (“Effective September 15, 2025”).

Version History

2025-11-11

Explicit AI clause: YES
Non-explicit predatory: NO
Flagged keywords: competing model ban, ai disclosure

Clause: competing model ban

Do Not Abuse our Platform. This includes using our products or services to: Coordinate malicious activity across multiple accounts to avoid detection or circumvent product guardrails or generating identical or similar inputs that otherwise violate our Usage Policy; Utilize automation in account creation or to engage in spammy behavior; Circumvent a ban through the use of a different account, such as the creation of a new account, use of an existing account, or providing access to a person or entity that was previously banned; Access or facilitate account or API access to Claude to persons, entities, or users in violation of our Supported Regions Policy; Intentionally bypass capabilities, restrictions, or guardrails established within our products for the purposes of instructing the model to produce harmful outputs (e.g., jailbreaking or prompt injection) without prior authorization from Anthropic; Utilization of inputs and outputs to train an AI model (e.g., “model scraping” or “model distillation”) without prior authorization from Anthropic.

Clause: ai disclosure

All consumer-facing chatbots, including any external-facing or interactive AI agent, must disclose to users that they are interacting with AI rather than a human. This disclosure must be provided at a minimum at the beginning of each chat session. […] Disclosure: If model outputs are presented directly to individuals or consumers, you must disclose to them that you are using AI to help produce your advice, decisions, or recommendations. This disclosure must be provided at a minimum at the beginning of each session.