Neue Forschungsergebnisse deuten darauf hin, dass die Beratung durch künstliche Intelligenz (KI) die Beziehungen zu menschlichen Fachkräften unbeabsichtigt belasten kann.
Using AI to verify human advice could damage your professional relationships
Neue Forschungsergebnisse deuten darauf hin, dass die Beratung durch künstliche Intelligenz (KI) die Beziehungen zu menschlichen Fachkräften unbeabsichtigt belasten kann.
Using AI to verify human advice could damage your professional relationships
5 Kommentare
100%
The amount of people who take it as Gospel is staggering.
EDIT: well, they believe it without question… but, generally, if you’re an expert in the field and you take the time to explain it (how llms gather information and how and what „they decide“ to answer) in a non-confrontational manner, most people can get over what the LLM told them.
It’s no different from the strain on relationships when one person says, „Hmmm… You may be right, but I’m going to ask these other people what they think“ or „I’m going to check a couple of the stack sites to see other people’s solutions.“
No one likes to be second-guessed.
*Everyone* should be second-guessed and get used to it.
Of course, when people complain about „no one uses critical thinking anymore“, what they mean is, „No one uses critical thinking in the way I use critical thinking.“
As designed, the paper cannot tell whether advisers react badly to AI in particular or to being cross-checked against a source they see as inferior, inappropriate, or low-status.
I would be interested to know how advisors would feel about clients getting advice from, for example, a family member. A family member may be seen as much less expert than a professional, but also isn’t necessarily competition in the same way that another advisor (or AI product) is. Plus, asking family for their opinions is normalized in a way that asking AI is not (yet). So I’ve no idea what the results would be, so it would be interesting.
Yes, it’s very embarrassing for professionals when AI knows vastly more than them, and is more accurate even with hallucinations. They’re seeing themselves become obsolete in real time.
If you’re consulting with a professional to begin with, why are you getting second opinions from a clanker?