[GoLUG] The Strange Behavior of LLMs in Hiring Decisions: Systemic Gender and Positional Biases in Candidate Selection

[Kevin Chadwick]

On 24/05/2025 06:41, Steve Litt wrote:
> But right now they can probably be made safer than 95% of human
> drivers, meaning driverless cars will decrease accidents and other
> mishaps.

There is a big difference between LLMs and a radar and camera system that may
well not be relying much on AI. If you mean Teslas fully self driving
(supervised) camera only AI then that is not safe at all. Elon lies every day
and it has crashed a car and done things that even the worst drivers would not do.

"https://dawnproject.com/"

An AI might be less likely to take dangerous shortcuts that many humans take
like speeding or driving through a red light or illegally driving over flat
roundabouts. I don't know if you have those in the US but some roundabouts are
so small in Britain that Artics/Semis have to be able to drive over them and
some people in Britain even turn right at a flat roundabout!

It might be that those people decide not to use self driving though anyway.