Why waste billions of kilo joules of energy running AI systems for that, when you'll get legions of dirt cheap technical labor in the developing world, who'll do it for you for far less and at massive scale, with better acerbic language?
I think part of the problem is that LLMs seem to be quite effective at producing messages adhering to ulterior motives, catch attention, reinforce emotions etc.
The GPT-4 release documentation has some examples of this in its addendum. ChatGPT also seems to be good at writing advertisements. Without the strong guardrails, I wouldn't bet on one or two persons instructimg a GPT-4-scale model perfoming worse at manipulating debates than 10 or 100 humans without AI.
reply