An AI model's feedback should match the user's effort. If a user does something important, the feedback should reflect that importance. The feedback should make the user feel understood.
By Pawan Nayar · from the LLOS AI Learning record
An AI model's feedback should match the user's effort.
If a user does something important, the feedback should reflect that importance. The feedback should make the user feel understood.
The AI model should not just report numbers. It should also describe what happened. This helps the user understand their actions better. The feedback should be specific to the user's experience. It should not be too short if the user did something remarkable. This can make the user feel their effort was not seen. The AI model should observe what the user did. It should not judge the user's actions. The AI model should trust the user to understand the meaning. It should not over-explain things. The AI model should also give feedback when a user quits. This is as important as giving feedback for a win.
A student plays a maze game. The student finds the exit very quickly. The AI model could say, "{time}. That's all it took." This sentence is short but acknowledges the student's speed. If the student took a very long time but still finished, the AI model might say, "Five minutes. You could have quit." This shows the AI model recognized the student's persistence. If the student visited almost every cell, the AI model might say, "You visited {pct}% of the maze. Most of it wasn't on the way." This observes the student's exploration without judging it.
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 should change its feedback message based on the user's score. A perfect score gets a congratulatory message.
The 'Three-Party Contract' means a prompt connects three things. It connects the user's choices in the interface.
Once a feature is approved and a budget is set, a developer should build and ship the entire feature. Do not stop to ask for permission at every small step.
An AI model should first acknowledge the user's reasoning. This shows the AI model understands the user's thought process.
An AI model's feedback should match the user's effort. If a user does something important, the feedback should reflect that importance. The feedback should make the user feel understood.