1. What the document is: Mistral AI's "Usage Policy" for Le Chat and Mistral AI Studio — a prohibited-content/conduct policy. Explicitly does not apply to Mistral AI products deployed on customer infrastructure, partner infrastructure, or Mistral's open-source models.

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.

Le Chat – 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/Le Chat/Acceptable Use Policy/ (raw Markdown captures, Open Terms Archive genai-eu dataset — no pre-tagged JSONL exists for this source) Last updated: 2026-03-26

Note on methodology: No pre-computed risk_score/keyword fields exist for this source. Keyword tags on this page are LLM-assigned. 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.

Data-quality note: Both raw captures for this doc type contain the entire policy text twice within the same file (a confirmed scraper/page-rendering artifact present in both dates, not a new issue introduced by either edit) — the quotes below are taken from the first occurrence only.

Overview

1. What the document is: Mistral AI’s “Usage Policy” for Le Chat and Mistral AI Studio — a prohibited-content/conduct policy. Explicitly does not apply to Mistral AI products deployed on customer infrastructure, partner infrastructure, or Mistral’s open-source models.

2. Input rights: Not addressed — purely a conduct policy.

3. Output restrictions: None in the rights/licensing sense — only content-type prohibitions (CSAM, non-consensual intimate imagery, hate speech, self-harm content, fraud, misinformation, unlicensed professional advice, privacy-invading impersonation via “someone else’s likeness or voice”).

4. Non-explicit predatory clauses: None found — no Feedback clause, no “improve services” language, and notably no AI-disclosure requirement at all (unlike OpenAI’s, Anthropic’s, and DeepSeek’s equivalent policies) — a gap worth noting in the cross-provider comparison.

5. Regulatory references: None found.

6. Regional variation: Not addressed in this document.

7. Key risk to users: Low. This is a standard prohibited-use list. The explicit “does not apply to… infrastructure deployed by a customer” carve-out is worth tracking, since it means self-hosted/on-prem Mistral deployments are governed by a different (uncaptured) agreement.

No AI-training, data-licensing, feedback-rights, or AI-disclosure clauses found.

Regulatory & Research Context

Pandit et al. (2026), who directly coded Mistral/Le Chat’s terms in their FAccT analysis, observe that “consumers suffer from lack of necessary information, significant imbalance of power” under GenAI terms generally — a dynamic this page illustrates structurally, since the policy’s complete silence on AI-disclosure (noted in Overview point 4 as a gap relative to OpenAI’s, Anthropic’s, and DeepSeek’s equivalents) leaves users with less information about AI involvement than peer providers supply. Applying Davidson et al.’s (2026) “regulatory gray areas” framework by analogy (Le Chat was not among the providers they directly studied), the explicit carve-out for “infrastructure deployed by a customer” — which this policy does not govern — is exactly the kind of jurisdiction/scope ambiguity their framework is designed to surface.

Changes Summary

Date	What changed
2026-01-12	Baseline version (“Effective: November 28, 2025”).
2026-03-26	Genuine additions: the Privacy clause gained an explicit example (“using someone else’s likeness or voice to generate outputs or impersonate them, without their prior consent”); a new “Security Violations” clause added, banning attempts to compromise Mistral’s/third parties’ security, including circumventing “security protections and AI safety filters.”

Version History

2026-01-12

Explicit AI clause: YES
Non-explicit predatory: NO
Flagged keywords: (none)

No AI-training, data-licensing, feedback-rights, or AI-disclosure clauses found in this document — it is purely a content/conduct policy.

2026-03-26

Explicit AI clause: YES
Non-explicit predatory: NO
Flagged keywords: (none)

Still no flaggable rights/licensing clauses. Genuine content additions this date (Privacy-clause example, new Security Violations clause) are conduct restrictions, not within this wiki’s tracked keyword categories.

Clause: new addition (not separately concept-tagged) — Security Violations

Security Violations. You shall not use the Mistral AI Products to compromise, or attempt to compromise, the security of Mistral AI, the Mistral AI Products, or any other third party. This includes creating malware and exploiting vulnerabilities. You shall not try to circumvent security protections and AI safety filters.