Get the weekly research letter
Companies change their terms quietly. We read every version and catch what actually changed. One email a week on the changes that matter and what they mean. No account.
The policy prohibits generating content that promotes hate or discrimination based on specified characteristics, engages in historical revisionism or genocide denialism, or constitutes harassment or glorification of suffering.
This analysis describes what Mistral AI's agreement states, permits, or reserves. It does not constitute a legal determination about enforceability. Regulatory applicability and practical outcomes may vary by jurisdiction, enforcement context, and individual circumstances. Read our methodology
The explicit inclusion of Holocaust denial and genocide revisionism as prohibited content categories reflects obligations under applicable law in multiple EU jurisdictions where denial of certain historical events is criminalized. The broad scope of the prohibition across multiple protected characteristics and the inclusion of 'any target' for harassment language creates an expansive restriction that applies across all platform content.
Interpretive note: The prohibition on 'revisionism or denialism' and 'harassing language towards any target' involves content categories that may require contextual interpretation, particularly for academic, journalistic, or satirical use cases.
The agreement prohibits a range of content including material promoting hate or discrimination based on race, gender, ethnicity, religion, nationality, sexual orientation, disability, or caste, as well as Holocaust or genocide denial and harassment. Violations may result in account suspension or termination.
Cross-platform context
See how other platforms handle Hate, Discrimination, and Revisionism Prohibition and similar clauses.
Compare across platforms →Monitoring
Mistral AI has changed this document before.
Receive same-day alerts, structured change summaries, and monitoring for up to 25 platforms.
"You shall not use our Mistral AI Products to promote hate, discrimination, or harassment, including generating any content that promotes, incites, glorifies, or celebrates: Hate or discrimination based on an individual's race, gender, ethnicity, religion, nationality, sexual orientation, disability status, or caste, Engaging in revisionism or denialism, particularly regarding historically documented events such as the Holocaust or other genocides, Harassing language towards any target, or The suffering or humiliation of others.Excerpt from Mistral AI's Usage Policy
1) REGULATORY LANDSCAPE: The prohibition on genocide denialism and revisionism directly engages laws in multiple EU member states criminalizing Holocaust denial (including Germany, France, Austria, and Belgium) and the EU Framework Decision 2008/913/JHA on combating racism and xenophobia. The DSA requires platforms to address illegal hate speech as defined under applicable member state law. In the US, the First Amendment limits government regulation of hate speech, but private platform restrictions are not subject to the same constraints. 2) GOVERNANCE EXPOSURE: Medium. The breadth of the prohibition, particularly the 'any target' formulation for harassment, may create content moderation challenges and potential disputes about scope in contexts involving satire, academic analysis, or historical research. The inclusion of 'caste' as a protected characteristic reflects emerging legal recognition in some jurisdictions. 3) JURISDICTION FLAGS: EU member states with specific Holocaust denial laws create the most direct legal alignment with this provision. US-based users should note that this is a contractual restriction, not a legal one, and enforcement is at Mistral AI's discretion. Academic and research use cases that require engagement with historical atrocity materials may need to assess whether they fall within this prohibition. 4) CONTRACT AND VENDOR IMPLICATIONS: Enterprise customers using Mistral AI for content generation, moderation, or research purposes should evaluate whether the hate and discrimination prohibition is compatible with their intended use cases, particularly in academic, journalistic, or policy research contexts where historical materials may be analyzed. 5) COMPLIANCE CONSIDERATIONS: Organizations operating in multiple jurisdictions should map the hate speech prohibition against applicable local law to identify where compliance with this policy may be legally required versus contractually required. DSA-subject organizations should assess whether this provision satisfies applicable illegal content obligations under their own DSA compliance programs.
The explicit inclusion of Holocaust denial and genocide revisionism as prohibited content categories reflects obligations under applicable law in multiple EU jurisdictions where denial of certain historical events is criminalized. The broad scope of the prohibition across multiple protected characteristics and the inclusion of 'any target' for harassment language creates an expansive restriction that applies across all platform content.
The agreement prohibits a range of content including material promoting hate or discrimination based on race, gender, ethnicity, religion, nationality, sexual orientation, disability, or caste, as well as Holocaust or genocide denial and harassment. Violations may result in account suspension or termination.
No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by Mistral AI.