USAi System Rules of Behavior
These rules may be updated periodically to remain relevant, accurate, and compliant. All responsibilities and expected behavior for users in your organization’s IT policies may apply to USAi as well; if you have questions about how the policies interact, reach out to the responsible service or staff office.
User Responsibilities
Assume accountability: You are responsible for all actions performed using your account. Do not share your credentials, API keys, or allow others to use your access.
Validate accuracy: AI is a tool to augment human work, not replace it. You are ultimately responsible for reviewing, validating, and taking ownership of any work product that incorporates AI-generated output. AI models can produce inaccurate, biased, or incomplete information (inaccurate responses sometimes referred to as “hallucinations”). Always critically evaluate and independently verify any factual claims or critical information generated by the Service before use or dissemination.
Expect user monitoring: All activities, inputs, and outputs on the Service are subject to logging, monitoring, and auditing by authorized personnel for security, operational, and compliance purposes.
Federal records management: Be aware that your interactions with the Service (e.g., chat conversations, API outputs) may constitute a Federal Record. You are responsible for managing these records and, as required, exporting them to an appropriate agency system in accordance with your agency’s records management policies.
Verify model suitability: Review model benchmarks and other materials to identify the right model for your specific use case. Explore multiple models to identify the best fit.
Utilize tokens responsibly: When using the API, please be mindful of token costs; be efficient and avoid unnecessary expenses.
Report incidents immediately: You must promptly report any suspected or confirmed security incidents to usai-security@usai.gov and your agency’s help desk / security operations center. Examples of incidents include, but are not limited to:
- A compromised user account or API key.
- Highly unusual or malicious system behavior.
Data and information handling: Use the minimum amount of data necessary to accomplish your task. When possible, de-identify, anonymize, or use synthetic data. Adhere to agency policy(ies) on inputting CUI, PII or other potentially sensitive information into the system.
Prohibited Uses
- Engage in any illegal activity or support illegal acts.
- Input, process, or store classified national security information on this Service.
- Generate, store, or transmit content that is sexually explicit, hateful, harassing, discriminatory, or otherwise abusive or offensive.
- Engage in commercial activity or private financial gain, or endorse any product or service.
- Intentionally bypass, defeat, or test the Service’s security controls, content filters, or safety mechanisms.
- Transmit malware or chain letters, or engage in unauthorized mass communications.
- Infringe upon copyrights, trademarks, or other intellectual property rights.
Rules of Behavior contact information
We welcome feedback regarding our Rules of Behavior. If you have any feedback or questions about these rules, contact support@usai.gov.