OpenAI Guidelines for Responsible and Safe AI Use

OpenAI has released a set of guidelines for the responsible and safe use of ChatGPT to ensure that the transformative capabilities of large language models (LLMs) are applied effectively and safely. These guidelines emphasize that while AI can enhance efficiency and creativity in knowledge work, human oversight and adherence to organizational policies are essential to mitigate risks associated with inaccuracies and bias.

Human Oversight and Fact-Checking

Human-in-the-loop verification is required for all important work because ChatGPT can produce inaccurate or outdated information. Since LLMs generate responses based on patterns in training data rather than a real-time understanding of truth, users must double-check critical facts against trusted sources.

To improve the accuracy of responses when recency is a priority, OpenAI recommends the following tools:

• Search: Enables ChatGPT to pull information from current sources.
• Deep Research: Provides a more intensive retrieval of up-to-date information.

Users are encouraged to verify citations, open provided links, and manually verify details before relying on AI-generated responses.

Professional Advice and Expert Review

ChatGPT is a conversational AI and not a licensed professional; therefore, it is not a replacement for qualified expert guidance. Users must seek professional review for decisions involving the following domains:

• Medical advice
• Legal advice
• Financial advice

Because AI responses are based on data patterns and may not be appropriate for a specific individual's unique situation, users should defer to qualified experts or their organization's internal policies before taking action in these areas.

Bias Mitigation and Critical Assessment

AI models may contain biases, making it necessary for users to review outputs carefully and assess conclusions critically. OpenAI identifies bias mitigation as an ongoing area of research and encourages users to report biased or unsafe replies using the thumbs-down button or the report flow to help improve model safety for all users.

Compliance, Transparency, and Consent

Responsible AI use requires adherence to both organizational and platform-level policies:

Policy Adherence

Users should prioritize their company's internal AI rules first, followed by OpenAI's Usage Policies.

Transparency

When employed in school or workplace settings that require disclosure, users should maintain conversation links or logs to provide a transparent record of how ChatGPT contributed to the final work.

Data Consent

Users must obtain explicit consent before sharing another person's voice or data. Specifically, when using "Record mode" (where enabled by workspace admins), users must ensure all participants agree to the recording and follow organizational policies regarding data capture.