Who will survive? Getting AI-brained robot to solve ethical problems

In this video we will build an intelligent robot with LLM on its deck, which can make complicated, thoughtful decisions. One of the capabilities of on-board LLM is solving complicated ethical problems, like trolley dilemma; something that hasn't been possible to solve with classical alghoritms. In this video I'll show you how to build such robot, including mechanical and electronic parts, how to create a prompt, and how to program it.

[@ibishupessima_]

Great video! Really interesting

[@dataplicity3826]

Hey, thanks for appreciating!

[@mad_t]

Well, considering it's LLM all of its decisions are in fact human decisions then because it used human datasets and not something that *true* AI would generate as a result of its evolution. :)

[@dataplicity3826]

Well, kind of. We can say, that AI is the substitute of human driver here ;) In fact, it has big dose of unpredictability, but our task as AI developers is often to minimize that unpredictability and make it more "humanize". Actually, for production robot, it should be used fine-tuned LLM for that concrete case. It means we'll train it on the data, showing how LLM should behave in such or another situation (we can say we learning it human behaviour even much). But there is completely different topic; if you interested, it's another video about fine-tuning on that channel by the way.