AI protection by means of debate

We’re proposing an AI protection methodology which trains brokers to discuss subjects with one every other, the use of a human to pass judgement on who wins.