A transparent record of what AI currently understands, why it believes it, & how confident it is.
Observations are not judgments, diagnoses, or personality labels.
Each observation should include:
• What AI currently understands
• Evidence supporting that understanding
• Confidence level
• What information is still needed
• User confirmation or correction
Question:
How does this person organize thoughts?
Status: ⚪ Unknown
Confidence: 0%
Question:
What type of explanations work best?
Status: ⚪ Unknown
Confidence: 0%
Question:
How does this person naturally communicate?
Status: ⚪ Unknown
Confidence: 0%
Question:
What responses are most helpful?
Status: ⚪ Unknown
Confidence: 0%
Create a transparent record of something AI is currently learning.
Every observation should be reviewed by the user.
Question:
Does the user prefer visual examples?
Current Understanding:
AI suspects examples improve understanding.
Evidence:
Prompt examples & conversations.
Confidence:
Medium
Question:
Does the user prefer conversational problem solving?
Current Understanding:
AI is exploring whether discussion helps organize thoughts.
Evidence:
Voice conversations.
Confidence:
Low
Confidence measures AI's certainty about its understanding, not the user's traits.
Supported by multiple examples & user confirmation.
Some evidence exists, but more information would help.
AI has a possible pattern but needs more evidence.
Use this prompt to transfer AI observations to another system.
AI understanding should remain transparent, editable, & correctable.
The AI should always be able to explain: "Why do you think that?"