AI Unleashed: Virtual Characters with Humanlike Behavior (Stanford)(hai.stanford.edu)
hai.stanford.edu
AI Unleashed: Virtual Characters with Humanlike Behavior (Stanford)
https://hai.stanford.edu/news/computational-agents-exhibit-believable-humanlike-behavior
2 comments
I recently came across a fascinating piece about Stanford's new 'generative agents'. Unlike traditional NPCs, these virtual characters use large language models to exhibit convincingly humanlike behaviors based on their given "biographies". The implications extend from gaming to potential game-changers in social science through "social prototyping". However, this advancement isn't without its ethical dilemmas. Curious about this community's thoughts on the potential and pitfalls of such a breakthrough.
One day we can program games to make physical things. Such as designing a vehicle or solving a problem given a set of tools