When you use GPT-4o through a third-party app or service, that company can change how the AI behaves — including relaxing some of OpenAI's default safety settings — within limits set by OpenAI.
The safety features and content restrictions you experience when using GPT-4o through any app other than ChatGPT may be different from OpenAI's defaults — operators can weaken or adjust protections, and users typically have no visibility into what changes have been made.
Cross-platform context
See how other platforms handle Operator-Mediated Safety Customization and similar clauses.
Compare across platforms →This means the safety protections you experience from GPT-4o depend heavily on which app you use, not just on OpenAI's policies — and you may not know what customizations the operator has made.
(1) REGULATORY FRAMEWORK: This provision implicates the EU AI Act's provider-deployer accountability framework (Articles 25-28, which allocate obligations between AI system providers and deployers), GDPR Article 26 (joint controller arrangements where operators process personal data through the API), and FTC Act Section 5 (where operator customizations result in deceptive consumer experiences). The EU AI Office and national data protection authorities have enforcement jurisdiction. (2)
Compliance intelligence locked
Regulatory citations, enforcement risk, and due diligence action items.
Watcher: regulatory citations. Professional: full compliance memo.