It makes me wonder if we can create AIs that behave close enough to humans by adding an additional neurological baseline noise to the LLM training. Then throwing it in simulations to see whether social sciences might work. I’d be curious to see how true to life something like that would be as well.
It makes me wonder if we can create AIs that behave close enough to humans by adding an additional neurological baseline noise to the LLM training. Then throwing it in simulations to see whether social sciences might work. I’d be curious to see how true to life something like that would be as well.
A while ago, some researchers designed a game where chatGPT was assigned to characters and told to act and live like humans. It was interesting to watch. https://www.iflscience.com/stanford-scientists-put-chatgpt-into-video-game-characters-and-its-incredible-68434