1. What the document is: "xAI Enterprise Terms of Service" — the business/API contract for the xAI API and Grok Business (X.AI LLC). Distinct from xAI's consumer Terms of Service.
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.
xAI – 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/xAI/Commercial Terms/ (raw Markdown captures, Open Terms Archive genai-eu dataset — no pre-tagged JSONL exists for this source)
Last updated: 2025-12-11 (content unchanged since baseline; see methodology note)
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 (train AI/models,competing model ban,feedback,input license). 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: “xAI Enterprise Terms of Service” — the business/API contract for the xAI API and Grok Business (X.AI LLC). Distinct from xAI’s consumer Terms of Service.
2. Input rights: The Customer keeps ownership of Input and Output. xAI gets only a purpose-limited license: “non-exclusive, worldwide, irrevocable, transferable, fully paid-up, royalty-free right and license to use, copy store, transmit, modify, and display the User Content in order to: (a) provide the Services… and (b) enforce xAI policies, to prevent abuse, and to perform… safety, compliance, and moderation.” There’s no “improve services” catch-all in this grant. Separately, there’s a strong, near-unconditional no-training commitment: “xAI shall not use any User Content for any of its internal AI or other training purposes (such as training its machine learning models), including developing new products or services based on User Content” — the dataset’s fourth such unconditional guarantee, after Llama API, Claude.ai Commercial Terms, and Perplexity Developer Terms. The one exception: xAI can still create and use de-identified data from Customer’s usage “to improve xAI’s products and services, to develop new products and services, and for its other business purposes.”
3. Output restrictions: “Customer shall not represent that Output was human-generated or use the Output to train Customer’s or its providers’ machine learning or AI models” — one sentence that combines an AI-disclosure duty with a competing-model restriction.
4. Non-explicit predatory clauses: The Feedback clause isn’t a license — it’s a full rights assignment: “Customer hereby assigns and shall cause all Permitted Users to assign to xAI all right, title and interest in and to such Feedback.” xAI is the dataset’s third provider (after Perplexity and Qwen Chat) to use assignment language instead of a license grant.
5. Regulatory references: No GDPR or EU AI Act citation appears in this document itself — those are handled via a separately-referenced DPA and Privacy Policy. It does disclose a HIPAA Protected Health Information restriction and a PCI DSS financial-data restriction on ordinary Input, each requiring a separate Business Associate Agreement — matching the pattern in ChatGPT’s Business Privacy Policy elsewhere in this dataset.
6. Regional variation: None. Governing law is Tennessee, but exclusive venue is Tarrant County, Texas — an internal inconsistency between the governing-law state and the venue state.
7. Key risk to users: Low-to-moderate. The purpose-limited Input/Output license and the strong no-training-on-raw-content commitment are both genuinely protective. The main risk is the de-identified-data carve-out, which has no specific limits on its scope, and the Feedback rights assignment.
Flagged Keywords & Risks (LLM-assigned)
train AI/models— “xAI shall not use any User Content for any of its internal AI or other training purposes… For clarity, subject to Section 3.2, xAI shall not use any User Content for any of its internal AI or other training purposes.” Why it matters: one of the strongest no-training commitments in the dataset — though de-identified data can still be used for general product improvement.competing model ban— “use any Service to help develop, or help provide to any third party, any product or service similar to or competitive with any Service,” plus “Customer shall not… use the Output to train Customer’s or its providers’ machine learning or AI models.” Why it matters: this bans both building a competing service and using Output to train any AI model — broader than a typical competing-product restriction.feedback— “Customer hereby assigns and shall cause all Permitted Users to assign to xAI all right, title and interest in and to such Feedback.” Why it matters: a full assignment — as strong as the Feedback clauses in Perplexity and Qwen Chat, the only other providers in this dataset that use assignment language.input license— purpose-limited to providing the Services and to safety/compliance/moderation, explicitly excluding training or product-improvement uses. Why it matters: one of the narrower, more protective Input/Output licenses in this dataset.
Regulatory & Research Context
xAI is one of the five providers Davidson et al. (2026) directly studied, though their specific findings on Grok concern consumer-facing misuse (non-consensual imagery) and jurisdictional clauses rather than this business-to-business document. Their broader concern with “regulatory gray areas” that create uncertainty for legitimate use applies by analogy to this page’s near-unconditional train AI/models no-training commitment — “xAI shall not use any User Content for any of its internal AI or other training purposes” — which, unlike the consumer Terms of Service, gives enterprise customers clearer ground rules, tempered only by the open-ended de-identified-data carve-out for “other business purposes.” Pandit et al. (2026) did not study xAI, but their observation that GenAI terms commonly extend Feedback rights without compensation applies by analogy to this document’s feedback clause, a full assignment “without providing any attribution or compensation.”
Changes Summary
| Date | What changed |
|---|---|
| 2025-11-11 | Baseline version (“Last Updated: June 27, 2025”). |
| 2025-12-02 | Cosmetic only: trailing-slash flicker on the top-of-page legal-hub link. |
| 2025-12-11 | Cosmetic only: trailing-slash flicker reverted. |
Version History
2025-11-11
- Explicit AI clause: YES
- Non-explicit predatory: YES
- Flagged keywords:
train AI/models,competing model ban,feedback,input license
Clause: train AI/models
xAI may create and use de-identified data related to Customer’s use of the Services to improve xAI’s products and services, to develop new products and services, and for its other business purposes (and such de-identified data will be owned by xAI). For clarity, subject to Section 3.2, xAI shall not use any User Content for any of its internal AI or other training purposes (such as training its machine learning models), including developing new products or services based on User Content.
Clause: competing model ban
(b) use any Service to help develop, or help provide to any third party, any product or service similar to or competitive with any Service… (e) scrape any User Content, distill model behavior, or remove or obscure any copyright or proprietary or other notice contained in any Service or Documentation.
Customer shall not represent that Output was human-generated or use the Output to train Customer’s or its providers’ machine learning or AI models.
Clause: feedback
In the event Customer or any Permitted User provides xAI with any suggestions, ideas, improvements or other feedback with respect to any aspect of the Services (“Feedback”), Customer hereby assigns and shall cause all Permitted Users to assign to xAI all right, title and interest in and to such Feedback, including all intellectual property rights therein, and acknowledges that xAI shall own such Feedback.
Clause: input license
Customer hereby grants to xAI a non-exclusive, worldwide, irrevocable, transferable (as set forth in Section 14.1), fully paid-up, royalty-free right and license to use, copy store, transmit, modify, and display the User Content in order to: (a) provide the Services to Customer; and (b) enforce xAI policies, to prevent abuse, and to perform operations relating to safety, compliance, and moderation.
2025-12-11
All flagged clauses unchanged from 2025-11-11 — only a cosmetic link-formatting flicker across both intervening captures.