- 1.The software cannot ask for clarification today because it doesn’t have a standardized format to request human inference as it is necessary.
- 2.It is possible to detect errors by querying large networks; it is possible to map these errors to underlying data gaps to direct learning. Mapping these differences requires a heterogeneous, large scale pool of individuals to request clarification from - HUMAN Protocol currently has over 100M individuals answering questions each month.
- 3.HUMAN Protocol defines contracts in software that manages both quality control and payments, interacting with humans who answer questions in 274 countries and territories.
Letting machines ask people for the questions they need to improve is the fundamental gateway to the next breakthrough in AI.