OpenAI, Google, and Anthropic enforce strict AI usage policies

Major AI platforms, including ChatGPT, Gemini, and Claude, maintain strict usage policies that can lead to account suspension or disabling. Recent incidents involving security researchers have highlighted the thin line between legitimate security research and policy violations.

One security researcher, known as log4j, reported that OpenAI disabled their ChatGPT account citing “Cyber Abuse” after the user published content regarding gaining administrative access to a 2025 Samsung TV to study data collection. OpenAI has not yet confirmed the specific actions that triggered the flag.

General grounds for account restrictions across these platforms include involvement in cyberattacks, fraud, impersonation, malware distribution, or privacy violations. Attempts to bypass safety mechanisms through jailbreaking or prompt injection are also prohibited. Additionally, platforms ban content related to child exploitation, terrorism, weapon manufacturing, and self-harm. Specific enforcement varies: OpenAI may disable accounts for sharing credentials or bypassing access limits, Google may restrict Gemini access for repeated violations, and Anthropic may lock Claude accounts for violating terms of service or accessing from unsupported regions.