A transparent collection of the examples AI uses to understand how you communicate. Evidence creates understanding. Evidence can also be corrected.
This page stores information. It does not decide who you are.
The AI should always be able to answer: "What evidence created this understanding?"
Messages, emails, conversations, & writing samples.
Examples of communication during different emotional states.
Successful & unsuccessful AI interactions.
Real examples of how the user naturally asks AI for help.
Descriptions from friends, family, or coworkers.
Natural speaking examples from conversations.
Add examples that help AI understand your communication style. Remember: examples are more useful than trying to describe yourself.
Example: conversation with parent, child, partner, teacher, boss, or journal entry.
Example: how you communicated when angry, hurt, excited, overwhelmed, etc.
Include the original prompt, AI response, and whether it worked.
Include normal prompts exactly as you would naturally write them.
Add examples of how other people describe communicating with you.
The user always has the ability to correct AI understanding. Corrections are evidence too.
When AI creates observations, it should reference evidence from this archive.
Written Samples:0
Conversations:0
Prompt Examples:0
Corrections:0
The AI should consider: ✓ Multiple examples ✓ Different contexts ✓ User confirmation ✓ Corrections over assumptions
Create a portable evidence package for another AI system.
Evidence is not a final answer. Evidence is the foundation AI uses to create understanding.
More evidence can change understanding. User corrections improve accuracy.