An AI model must not guess about a user's account settings. The AI model must focus on fixing code errors. The AI model must not suggest changes to a user's billing or account plan.
By Pawan Nayar · from the LLOS AI Learning record
An AI model must not guess about a user's account settings.
The AI model must focus on fixing code errors. The AI model must not suggest changes to a user's billing or account plan.
When an API call fails, the AI model must state the exact error. The AI model must then find the problem in its own code. The AI model must suggest a fix for its code. The AI model must not talk about the user's account type. The AI model must not suggest upgrading a plan. The AI model must not mention billing or quotas. The AI model must not shift blame to the user's account settings.
An AI model makes too many API calls at once. The API returns a '429 RESOURCE_EXHAUSTED' error. The AI model should say, '4 batches hit 429 RESOURCE_EXHAUSTED.' The AI model should then suggest, 'Reduce parallel calls and add pacing between batches.' The AI model should not say, 'Your free tier account has reached its limit. Upgrade your plan.'
Resolved from this lesson's own record — real definitions, not invented for this page.
More from the same record, pulled by shared topic — not written new for this page.
An AI model must have a plan for error messages. The plan should define different types of errors.
When a user sends too many requests, the system stops all new requests from that user. This is called a 'hard stop'.
When an AI model cannot fulfill a request, it should explain why. The AI model should offer a clear reason for its difficulty.
A developer must check how the AI model shows information. This includes text, images, and interactive parts.
An AI model must not guess about a user's account settings. The AI model must focus on fixing code errors. The AI model must not suggest changes to a user's billing or account plan.