HUMAN Protocol AI Whitepaper

In this Whitepaper, we delve into the difficulties of training and achieving machine comprehension, and explore how an automated HUMAN platform can facilitate the next expansion in AI capabilities: for machines to request the data needed to improve their algorithms

If you’re building machine learning models, could the software determine that it doesn't know what it doesn’t know? Why couldn't the AI ask directly for clarification? When we see an edge case or error, can we ask: “what is the logical fallacy or gap that has led to this error,” and use that information to ask massively distributed networks of people for clarification? Can we build a decentralized web-scale marketplace that solves these problems?