You cannot use Mistral AI's products to try to hack systems, create malicious software, or find ways to bypass the AI's built-in safety restrictions, including through prompt manipulation techniques.
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 prohibition on circumventing AI safety filters is operationally significant because it targets jailbreaking and prompt injection techniques commonly used to extract prohibited outputs, and violations could result in account termination.
Interpretive note: The absence of a security research safe harbor creates ambiguity about whether legitimate adversarial testing or red-teaming activities are prohibited under this provision.
Users who attempt to manipulate the AI into producing prohibited content through creative prompting or jailbreak techniques may have their accounts suspended or permanently terminated under this provision.
How other platforms handle this
Don't misuse Gemini Apps by trying to work around safety measures or policies we have put in place. For example, don't try to use Gemini Apps to generate content that our measures would otherwise prevent.
We believe that the responsible development and maintenance of advanced AI for the long-term benefit of humanity is our mission. Iterative deployment is a key part of our safety strategy. Deploying models incrementally allows us to learn from real-world use and make improvements before more powerful...
You may not use the Services to circumvent, disable, or otherwise interfere with safety-related features or restrictions of the Services, including content filtering mechanisms or usage restrictions.
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 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.— Excerpt from Mistral AI's Mistral AI Usage Policy
(1) REGULATORY LANDSCAPE: This provision engages the Computer Fraud and Abuse Act (CFAA) in the U.S. and the EU's Directive on Attacks Against Information Systems (2013/40/EU) for the computer security aspects; the prohibition on circumventing AI safety filters also engages the EU AI Act's requirements for providers of general-purpose AI models to implement technical and governance measures against misuse. The FTC has enforcement authority over deceptive or unfair practices involving cybersecurity. (2) GOVERNANCE EXPOSURE: Medium. The prohibition on security violations is standard across AI platform policies. The additional clause prohibiting circumvention of AI safety filters is more specific and may create enforcement complexity around distinguishing legitimate adversarial testing (e.g., by security researchers) from prohibited jailbreaking, given the policy does not carve out security research. (3) JURISDICTION FLAGS: The CFAA's application to terms-of-service violations has been subject to significant litigation in the U.S. (including Van Buren v. United States), and the policy's prohibition on circumvention attempts may not map cleanly onto CFAA liability in all contexts. In the EU, the Directive on Attacks Against Information Systems provides a clearer framework for unauthorized access but similarly does not resolve ambiguities around AI filter circumvention. (4) CONTRACT AND VENDOR IMPLICATIONS: Enterprise customers conducting AI red-teaming, adversarial robustness testing, or security research involving Mistral AI's platform should assess whether those activities fall within this prohibition and may want to seek written clarification or contractual carve-outs from Mistral AI before conducting such testing. (5) COMPLIANCE CONSIDERATIONS: Organizations with security research or penetration testing functions that use Mistral AI's platform should review this provision and consider whether their testing activities require a separate agreement with Mistral AI. The absence of a security research safe harbor in this policy is notable for enterprise security teams.
Full compliance analysis
Regulatory citations, enforcement risk, and due diligence action items.
Free: track 1 platform + weekly digest. Monitor: 25 platforms + same-day alerts. No credit card required.
Compliance Governance Intelligence
Need to monitor specific governance provisions?
Compliance includes provision-level monitoring, governance timelines, regulatory mapping, and audit-ready analysis.
Built from archived source documents, structured governance mappings, and historical version tracking.
The explicit prohibition on circumventing AI safety filters is operationally significant because it targets jailbreaking and prompt injection techniques commonly used to extract prohibited outputs, and violations could result in account termination.
Users who attempt to manipulate the AI into producing prohibited content through creative prompting or jailbreak techniques may have their accounts suspended or permanently terminated under this provision.
No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by Mistral AI.