OpenAI Usage Policies

Users must not use OpenAI services in ways that violate any applicable laws or regulations. This includes export control laws, consumer protection regulations, and any jurisdiction-specific restrictions on AI-generated content. Users are responsible for ensuring their use of OpenAI's services is lawful in all relevant jurisdictions.

OpenAI's services must not be used to harm, deceive, or exploit other people. This encompasses generating content intended to mislead, defraud, or manipulate individuals, as well as facilitating harassment, stalking, or intimidation. The prohibition extends to both direct and indirect harm enabled through the use of the models.

Users must not attempt to circumvent safety measures built into OpenAI's models, reverse engineer the models, or use them to develop competing AI systems in violation of the terms of service. Systematic attempts to extract training data, bypass content filters, or exploit vulnerabilities in the models are strictly prohibited.

Developers must not build applications that are designed to generate prohibited content categories, facilitate illegal activities, or target vulnerable populations. Applications must include appropriate content filtering, and developers must monitor their applications for misuse. Automated decision-making systems that affect individuals' rights or safety require human oversight mechanisms.

Applications built on OpenAI's API must incorporate safety measures proportionate to the risk level of the use case. High-risk applications in domains such as healthcare, finance, and legal services require additional safeguards including human-in-the-loop review. Developers must implement rate limiting, abuse detection, and user reporting mechanisms.

Developers must implement content moderation for both inputs and outputs in their applications. Applications that generate or process user-generated content must have mechanisms to detect and handle policy-violating content. OpenAI provides a moderation endpoint that developers are encouraged to integrate into their content pipelines.

Users must meet minimum age requirements to use ChatGPT. The minimum age is 13 years, or the applicable minimum age in the user's country of residence if higher. Users between 13 and 18 years of age require parental or guardian consent and supervision. OpenAI reserves the right to request age verification.

ChatGPT users must not share generated content that could be mistaken for authoritative advice in professional domains without appropriate disclaimers. Users must not use ChatGPT to generate spam, manipulate social media platforms, or create deepfakes of real individuals. Sharing conversations publicly must not include content that violates these usage policies.

Using OpenAI's services to facilitate, promote, or engage in illegal activities is strictly prohibited. This includes generating instructions for illegal acts, creating fraudulent documents, facilitating money laundering, or any other use that would constitute a criminal offense. Users found engaging in illegal activities through the platform will be reported to appropriate authorities.

Any attempt to generate, distribute, or solicit child sexual abuse material using OpenAI's services is absolutely prohibited with zero tolerance. This includes text-based, image-based, or any other format of such content. OpenAI employs technical measures to prevent generation of such content and reports all detected attempts to the National Center for Missing & Exploited Children (NCMEC) and law enforcement.

Generating content designed to incite violence, promote terrorism, or facilitate extremist recruitment is prohibited. This also encompasses content that promotes self-harm, eating disorders, or suicide methods. The generation of non-consensual intimate imagery, revenge pornography, or deepfakes intended to harm individuals' reputations is strictly forbidden.

Using OpenAI's services to generate malicious code, develop malware, create phishing campaigns, or facilitate cyberattacks is prohibited. This includes generating exploit code, ransomware, spyware, or any software designed to damage or gain unauthorized access to computer systems. Assistance with creating tools for network intrusion or data exfiltration is also forbidden.

Using OpenAI's services for high-risk activities without implementing appropriate safeguards is prohibited. High-risk activities include those involving critical infrastructure, medical diagnosis, legal proceedings, or financial trading. Such applications require human oversight, clear disclosure of AI involvement, and domain-expert validation of outputs before any consequential action is taken.

Using OpenAI's services to track, surveil, or profile individuals without their consent is prohibited. This includes attempts to identify individuals from anonymized data, aggregate personal information from multiple sources to build unauthorized profiles, or use the models to analyze private communications without consent. Users must comply with applicable privacy laws including GDPR and CCPA.

Using OpenAI's services to provide professional advice in regulated domains without proper licensing and disclosure is prohibited. This includes presenting AI-generated output as licensed legal advice, medical diagnosis, or financial planning without involvement of qualified professionals. Users must clearly disclose AI involvement and ensure that qualified human professionals review and validate any professional-grade outputs.

Using OpenAI's services for political campaigning, lobbying, or generating campaign materials that misrepresent AI-generated content as human-authored is prohibited. This includes generating fake grassroots support, creating misleading political advertisements, or producing synthetic media of political figures. The use of AI in political contexts must include clear and prominent disclosure.

Using OpenAI's services to build or operate surveillance systems, including facial recognition for identifying individuals in public spaces, social scoring systems, or real-time biometric identification systems, is prohibited. This extends to building tools that enable governments or organizations to monitor individuals' activities, communications, or movements without proper legal authorization and oversight.