- Reaction score
- 2,887
- Points
- 1,140
We have politicians who don't even know not to read the AI prompts.and even worse, it teaches things that are flat out wrong although they often come with the "this may not be exact" caveat. But who pays attention to a warning until they actually get burnt. Case in point I was out pulling weeds yesterday and came across one without any fruit or flowers that I couldn't readily identify as they were only a few inches above ground. So I asked THE MACHINE and it came back with young mock orange plants: good deal. But there was a vine growing nearby with a similar leaf so I asked about it as well. Came back with a plant called 'dog strangling vine' which is listed as noxious and invasive and can be toxic. So I went back to sample A and tried again. This time it came back as young vine plants. AI is not intelligent nor does it think for itself. It simply allows the user to access a much larger data base than a home computer or laptop and processes that larger resource to produce the most logical answer given the information. Believing it as being fact can and will lead to wrong conclusions and possible fatal errors. Could be of interest to game the 1983 false alarm regarding missile launches and again the 1995 Norwegian launch to see what an AI response would make of them. Would they agree with Petrov and ignore it instead of passing it up the line, would they wait as the Russians did in 1995? Do want to bet your life on a computer decision?
Sure its just a backbencher in a provincial legislature, but that's just because he outed himself.
How many cabinet ministers or premiers are using LLMs to make decisions or tell them how to think?

