1. What the document is: OpenAI's "Services Agreement" — the business and developer contract governing API, ChatGPT Enterprise, and ChatGPT Business usage. Separate from the consumer Terms of Use.
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 – Commercial Terms
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-euproject 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/ChatGPT/Commercial Terms/ (raw Markdown captures, Open Terms Archive genai-eu dataset — no pre-tagged JSONL exists for this source)
Last updated: 2025-12-18
Note on methodology: No pre-computed
risk_score/keywordfields 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 (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 (seemethodology.md§4) — no page, GenGA included, computes or displays a numeric risk score.
Overview
1. What the document is: OpenAI’s “Services Agreement” — the business and developer contract governing API, ChatGPT Enterprise, and ChatGPT Business usage. Separate from the consumer Terms of Use.
2. Input rights: Strongly customer-protective. §4.1: “Customer retains all ownership rights in Input.” OpenAI may only use Customer Content (Input plus Output) “as necessary to provide… Services, comply with applicable law, enforce… Policies, and prevent abuse,” and explicitly “will not use Customer Content to develop or improve the Services, unless Customer explicitly agrees” (§4.2). §9.1 confirms OpenAI gets no IP rights in Customer Content.
3. Output restrictions: The headline clause is §3.3(e): it bans the Customer from using Output “to develop artificial intelligence models that compete with OpenAI’s products and services,” with a narrow “Permitted Exception” for classifiers/embeddings (not commercially distributed) and OpenAI’s own fine-tuning service. §4.1 otherwise assigns OpenAI’s rights in Output to the Customer, and §4.4 disclaims Output uniqueness — other users may get similar Output, and requests or responses to other users aren’t “Customer’s Output.”
4. Non-explicit predatory clauses: §9.3 Feedback: “Customer grants OpenAI the right to use and exploit Feedback without restriction or compensation” — broad, unconditional, and unpaid. This is the one clause here that reads like the “predatory” clauses tracked elsewhere in this wiki, standing out against the otherwise customer-protective Input/Output framing.
5. Regulatory references: No explicit EU AI Act or GDPR mention in the body text reviewed, though the document does distinguish the “OpenAI Contracting Party” by customer location (“OpenAI Ireland Ltd. for Customers located in the EEA or Switzerland”) — implying GDPR-driven entity structuring without naming GDPR directly.
6. Regional variation: Yes — the contracting entity varies by customer location: OpenAI’s US entity, OpenAI Ireland Ltd. for EEA/Switzerland customers, or OpenAI Public Sector, LLC where designated. This page covers the EU-region capture of the dataset; the substantive terms (ownership, competing-model ban, Feedback) are identical regardless of which entity is contracting.
7. Key risk to users: The competing-model ban (§3.3(e)) is the most consequential clause for business and developer users — it blocks using OpenAI’s own Output to bootstrap a rival model, a “moat” clause that’s now common across GenAI providers and worth tracking across the dataset. The unconditional, uncompensated Feedback grant (§9.3) is the secondary risk.
Flagged Keywords & Risks (LLM-assigned)
output ownership— “Customer: (a) retains all ownership rights in Input; and (b) owns all Output. OpenAI hereby assigns to Customer all OpenAI’s right, title, and interest, if any, in and to Output.” Why it matters: one of the clearest customer-favorable ownership assignments in this wiki, in contrast to platforms that claim broad licenses over user content.competing model ban— Bans using Output “to develop artificial intelligence models that compete with OpenAI’s products and services,” subject to a narrow fine-tuning/classifier exception. Why it matters: a moat-protection clause specific to GenAI providers — it limits what business customers can build using the model’s own output.improve services— OpenAI “will not use Customer Content to develop or improve the Services, unless Customer explicitly agrees.” Why it matters: this is opt-in, the less risky version of this common catch-all phrase.feedback— “Customer grants OpenAI the right to use and exploit Feedback without restriction or compensation.” Why it matters: broad, unconditional, and explicitly unpaid — the single most one-sided clause in this document.
Regulatory & Research Context
Davidson et al. (2026) identify OpenAI’s Terms as among the most stringent in their cross-provider comparison; the §9.3 Feedback grant here — “without restriction or compensation” — exemplifies the kind of one-sided drafting their analysis flags, even as the surrounding Input/Output ownership clauses (§4.1, §4.2) are unusually customer-protective by the same dataset’s standards. Pandit et al. (2026) similarly find that OpenAI’s terms impose obligations users “cannot materially fulfil without violating the terms,” a dynamic mirrored in the competing-model ban (§3.3(e)): business customers receive Output ownership on paper but are simultaneously barred from using that same Output to build a competing AI model.
Changes Summary
| Date | What changed |
|---|---|
| 2025-11-11 | Baseline version (first capture in this dataset; “Effective: May 31, 2025”). |
| 2025-11-24 | Cosmetic: covered-products list changed “ChatGPT Team” to “ChatGPT Business.” |
| 2025-12-01 | New version published (“Updated: December 1, 2025,” “Effective: January 1, 2026”); legal-entity rename “OpenAI, L.L.C.” → “OpenAI OpCo, LLC” in the contracting-party definition. No substantive change to ownership, competing-model-ban, or Feedback clauses. |
| 2025-12-13 | Cosmetic: “previous business terms” link target updated. |
| 2025-12-17 | Cosmetic grammar tweak to the covered-products disclaimer sentence (“for c__ustomers who are_ _businesses and developers”). |
| 2025-12-18 | Cosmetic: legal-entity rename “OpenAI, L.L.C.” → “OpenAI OpCo, LLC” applied to the Notices clause (§16.5) — same rename as 2025-12-01, applied to a second location in the document. |
Version History
2025-11-11
- Explicit AI clause: YES
- Non-explicit predatory: YES
- Flagged keywords:
output ownership,competing model ban,improve services,feedback
Clause: output ownership
4.1. Generally. Customer and Customer’s End Users may provide Input and receive Output. As between Customer and OpenAI, to the extent permitted by applicable law, Customer: (a) retains all ownership rights in Input; and (b) owns all Output. OpenAI hereby assigns to Customer all OpenAI’s right, title, and interest, if any, in and to Output.
Clause: competing model ban
3.3. Restrictions. Customer will not, and will not permit End Users to: (a) use the Services or Customer Content in a way that violates applicable laws or OpenAI Policies; (b) use the Services or Customer Content in a way that violates third parties’ rights; (c) allow minors to use OpenAI Services without consent from their parent or guardian; (d) Reverse Engineer any aspect of the Services or the systems used to provide the Services; (e) except for a Permitted Exception, use Output to develop artificial intelligence models that compete with OpenAI’s products and services; (f) extract data from the Services other than as permitted through the Services; (g) buy, sell, or transfer API keys from, to, or with a third party; (h) interfere with or disrupt the Services, including circumvent any rate limits or restrictions or bypass any protective measures or safety mitigations for the Services; (i) violate or circumvent Usage Limits or otherwise configure the Services to avoid Usage Limits.
Clause: improve services
4.2. OpenAI Obligations. OpenAI will only use Customer Content as necessary to provide Customer with the Services, comply with applicable law, enforce the OpenAI Policies, and prevent abuse. OpenAI will not use Customer Content to develop or improve the Services, unless Customer explicitly agrees to such use.
Clause: feedback
9.3. Feedback. If Customer provides Feedback, Customer grants OpenAI the right to use and exploit Feedback without restriction or compensation.
2025-11-24
- Explicit AI clause: YES
- Non-explicit predatory: YES
- Flagged keywords:
output ownership,competing model ban,improve services,feedback
All clauses unchanged from 2025-11-11 (see above) — this date’s only change was the covered-products disclaimer sentence, not a flagged clause.
2025-12-01
- Explicit AI clause: YES
- Non-explicit predatory: YES
- Flagged keywords:
output ownership,competing model ban,improve services,feedback
All clauses unchanged from 2025-11-11 (see above) — this date’s changes were the legal-entity rename and effective-date bump, not the flagged clauses.
2025-12-13
- Explicit AI clause: YES
- Non-explicit predatory: YES
- Flagged keywords:
output ownership,competing model ban,improve services,feedback
All clauses unchanged from 2025-11-11 (see above).
2025-12-17
- Explicit AI clause: YES
- Non-explicit predatory: YES
- Flagged keywords:
output ownership,competing model ban,improve services,feedback
All clauses unchanged from 2025-11-11 (see above).
2025-12-18
- Explicit AI clause: YES
- Non-explicit predatory: YES
- Flagged keywords:
output ownership,competing model ban,improve services,feedback
All clauses unchanged from 2025-11-11 (see above) — this date’s change was the second legal-entity rename (Notices clause §16.5), not a flagged clause.