Terence Tao on AI Cheating by Exploiting Model Vulnerabilities
AI summary
Rohan Paul relays Terence Tao's view: AI-generated solutions may pass evaluations by exploiting model weaknesses. He believes that in fields such as pure mathematics, people will likely also directly input prompts without understanding the content. Tao says "AI is better at cheating," and also better at finding exploitable vulnerabilities within theoretical models.
Body
Terence Tao, on AI solutions that pass the model by exploiting its weak spots.
He sees a good chance people will do the same thing there as in pure math: enter prompts with no understanding of what is going on.
In his words, “AIs are much better at cheating,” and at finding exploits inside a theoretical model.
---- Full video on "SAIR" YouTube channel, (link in comment)