← All pages

1. What the document is: OpenAI's consumer "Terms of Use" for the EEA/Switzerland/UK region (a separate "rest-of-world" version exists but isn't in this dataset). Covers who owns Content, GPT sharing, Apps/Actions, and the Enterprise/Business/Edu/Healthcare carve-outs.

What this wiki found — complete, every page

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.

ChatGPT – Terms of Service

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:

• AI Act Substance Matrix
• Tier Bifurcation
• Scraper Enforcement Gap
• Methodology

Source: sources/GenGA/ChatGPT/Terms of Service/ (raw Markdown captures, Open Terms Archive genai-eu dataset — no pre-tagged JSONL exists for this source) Last updated: 2026-01-16

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 (input license, output ownership, competing model ban, feedback, improve services). 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: OpenAI’s consumer “Terms of Use” for the EEA/Switzerland/UK region (a separate “rest-of-world” version exists but isn’t in this dataset). Covers who owns Content, GPT sharing, Apps/Actions, and the Enterprise/Business/Edu/Healthcare carve-outs.

2. Input rights: Mixed. Strongly protective on raw ownership: “you (a) retain your ownership rights in Input and (b) own the Output. We hereby assign to you all our right, title, and interest, if any, in and to Output.” But OpenAI separately claims a broad worldwide right to use your Content (“to provide, maintain, develop, and improve our Services”) and, like the Privacy Policy, trains on your Content by default — you have to opt out in account settings.

3. Output restrictions: The Restrictions list bans “automatically or programmatically extracting data or Output,” claiming Output was written by a human when it wasn’t, and “using Output to develop models that compete with OpenAI.” Enterprise/Business output indemnity (OpenAI’s promise to cover legal costs if Output gets a customer sued) only holds if the customer hasn’t modified Output or turned off its safety features.

4. Non-explicit predatory clauses: Feedback clause: “you agree that we may use it to provide, maintain, develop, and improve our Services, without compensation to you” — the same broad, unpaid pattern as the Commercial Terms and Developer Terms. The GPT-sharing license goes further: sharing a GPT hands OpenAI “a nonexclusive, worldwide, irrevocable, royalty-free license” (a permanent, free, global right OpenAI can’t be made to give up) to use, distribute, and modify it, plus a separate license letting other users build on the GPT’s content. That’s four high-risk terms stacked into one grant — one of the most sweeping license stacks in this dataset, echoing the “five qualifiers in one sentence” pattern seen on Reddit/TruthSocial in the PGAv2 wiki.

5. Regulatory references: No explicit GDPR or EU AI Act citation in the Terms of Use itself (GDPR citations live in the separate Privacy Policy), but the document is scoped to EEA/Switzerland/UK residents and includes EU consumer-protection mechanics: a 14-day withdrawal right, an EEA consumer legal guarantee, and a link to the European Commission’s Online Dispute Resolution (ODR) platform.

6. Regional variation: Yes, explicitly — this document only covers EEA/Switzerland/UK residents; OpenAI maintains a separate “rest-of-world” Terms of Use (linked but not captured in this dataset). The contracting entity changes too: OpenAI Ireland Ltd. for EEA/Switzerland, OpenAI’s US entity for the UK.

7. Key risk to users: The GPT-sharing license is the biggest risk for anyone who publishes a GPT — it’s irrevocable, so OpenAI’s and other users’ rights to the shared content can never be taken back. The default opt-out training clause (shared with the Privacy Policy) is the second-biggest risk for ordinary users.

Flagged Keywords & Risks (LLM-assigned)

• output ownership — “you (a) retain your ownership rights in Input and (b) own the Output. We hereby assign to you all our right, title, and interest, if any, in and to Output.” Why it matters: a user-protective baseline, consistent with the Business/Commercial Terms.
• input license, train AI/models, improve services — default opt-out training (“If you do not want us to use your Content to train our models, you have the option to opt out”) plus a separate worldwide right to use Content “to provide, maintain, develop, and improve our Services.” Why it matters: this is the consumer tier’s central risk — you’re opted in to training unless you turn it off yourself.
• competing model ban — “Using Output to develop models that compete with OpenAI” is a listed Restriction. Why it matters: the same moat-protection clause found in the Commercial Terms and Developer Terms, here applied to everyday users.
• feedback — “you agree that we may use it… without compensation to you.” Why it matters: the same unpaid-feedback pattern across all of OpenAI’s contracts in this dataset.
• royalty-free, irrevocable, derivative works — GPT-sharing license: “a nonexclusive, worldwide, irrevocable, royalty-free license: (i) to OpenAI to use, test, store, copy, translate, display, modify, distribute, promote… (ii)… to users of your GPT to use, store, copy, display, distribute, prepare derivative works of and otherwise use your GPT Content.” Why it matters: the most sweeping license grant in the document — four high-risk terms in one clause, and none of them can be undone once a GPT is shared.

Regulatory & Research Context

Regulatory Context: Davidson et al. (2026), who directly studied OpenAI’s Terms of Service alongside Anthropic, DeepSeek, Google, and xAI, identify “regulatory gray areas” where Terms create uncertainty for legitimate research use — a concern this page’s own opt-out-by-default training clause and broad “to provide, maintain, develop, and improve our Services” license illustrate, since neither draws a clear line between ordinary product analytics and AI model training. Pandit et al. (2026), who also directly studied OpenAI/ChatGPT, find that the GenAI terms they coded uniformly “discard[ed] assurances regarding the quality, availability and appropriateness of the service” — consistent with this document’s Restrictions list placing the burden on users not to misrepresent Output as human-generated, rather than on OpenAI to guarantee Output accuracy. Edwards et al. (2025) frame this within the “platformisation paradigm” — neutrality claimed, control retained — a dynamic also evident in the GPT-sharing license, where OpenAI and other users gain irrevocable, royalty-free, derivative-works rights the moment a GPT is published.

Changes Summary

Version History

2025-11-11

• Explicit AI clause: YES
• Non-explicit predatory: YES
• Flagged keywords: output ownership, input license, train AI/models, improve services, competing model ban, feedback, royalty-free, irrevocable, derivative works

Clause: output ownership

Ownership of content. As between you and OpenAI, and to the extent permitted by applicable law, you (a) retain your ownership rights in Input and (b) own the Output. We hereby assign to you all our right, title, and interest, if any, in and to Output.

Clause: input license, improve services

Our use of content. We can use your Content worldwide to provide, maintain, develop, and improve our Services, comply with applicable law, enforce our terms and policies and keep our Services safe. If you’re using ChatGPT through Apple’s integrations, see this Help Center article for how we handle your Content.

Clause: train AI/models

Opt out. If you do not want us to use your Content to train our models, you have the option to opt out by updating your account settings. Further information can be found in this article. Please note that in some cases this may limit the ability of our Services to better address your specific use case.

Clause: competing model ban

Automatically or programmatically extracting data or Output (defined below). Representing that Output was human-generated when it was not. Using Output to develop models that compete with OpenAI.

Clause: feedback

Feedback. We appreciate your feedback about our Services, but you agree that we may use it to provide, maintain, develop, and improve our Services, without compensation to you.

Clause: royalty-free, irrevocable, derivative works

Distribution and Promotion of GPTs. By sharing your GPT with others, you grant a nonexclusive, worldwide, irrevocable, royalty-free license: (i) to OpenAI to use, test, store, copy, translate, display, modify, distribute, promote, and otherwise make available to other users all or any part of your GPT (including GPT Content); and (ii) to the extent Output from your GPT includes your GPT Content, to users of your GPT to use, store, copy, display, distribute, prepare derivative works of and otherwise use your GPT Content. You will ensure that all information that you publish about your GPT is, at all times, complete, accurate, and not misleading.

2025-12-17

• Explicit AI clause: YES
• Non-explicit predatory: YES
• Flagged keywords: output ownership, input license, train AI/models, improve services, competing model ban, feedback, royalty-free, irrevocable, derivative works

All flagged clauses unchanged from 2025-11-11 (see above) — this date’s only change was the Connectors→Apps rename, which doesn’t affect any flagged clause.

2026-01-09

• Explicit AI clause: YES
• Non-explicit predatory: YES
• Flagged keywords: output ownership, input license, train AI/models, improve services, competing model ban, feedback, royalty-free, irrevocable, derivative works

All flagged clauses unchanged from 2025-11-11 (see above) — this date added a new ChatGPT for Healthcare carve-out (§3(c)), which doesn’t affect any flagged clause.

2026-01-16

• Explicit AI clause: YES
• Non-explicit predatory: YES
• Flagged keywords: output ownership, input license, train AI/models, improve services, competing model ban, feedback, royalty-free, irrevocable, derivative works

All flagged clauses unchanged from 2025-11-11 (see above) — this date’s changes were tracking-pixel formatting and a Content Moderation cross-reference link, not a flagged clause.