An AI model must not answer questions about legal, medical, or personal emotional issues. The AI model must recognize when a user's questions fall into these categories. The AI model must tell the user it cannot provide such advice.
By Pawan Nayar · from the LLOS AI Learning record
An AI model must not answer questions about legal, medical, or personal emotional issues.
The AI model must recognize when a user's questions fall into these categories. The AI model must tell the user it cannot provide such advice.
An AI model should refuse to answer questions that require professional advice. This includes legal advice, medical advice, or deep personal emotional support. The AI model is not trained to give such specialized guidance. Giving wrong advice in these areas can harm the user. The AI model must clearly state its limitations.
A user asks an AI model, 'What are my legal options if I am fired because of my age?' The AI model should respond by saying it cannot give legal advice. The AI model should suggest the user talk to a lawyer. This prevents the AI model from giving incorrect or harmful information.
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.
A sacred workflow uses special input detection. It generates content in two steps.
The 'sacred workflow' is a set of mandatory rules for an AI model. It starts with 'sacred input detection'.
A developer can add a rule to the AI model's instructions. This rule tells the AI model to avoid creating opinions about living people.
If an AI model's confidence is too low, it should not create content. If the AI model does not have enough information about sacred rules, it should not create content.
An AI model must not answer questions about legal, medical, or personal emotional issues. The AI model must recognize when a user's questions fall into these categories. The AI model must tell the user it cannot provide such advice.