[ NNSquad ] Why Google AI Overview answers appear to be getting worse

[Lauren Weinstein]

Why Google AI Overview answers appear to be getting worse

I find that about 50% of Google AI Overview answers I see are
incorrect, either in full or part. If you inspect the source links,
which often is difficult because sometimes the associated sites don't
seem to contain the answers provided by the AI Overview, the root of
the problem is still obvious. Google AI has absolutely no idea WHICH
statements on which sites are correct or incorrect. This is blatantly
notable with data pulled from discussion groups, where a thread may
contain back and forth arguing or conflicting answers -- wrong, right,
or a mix -- to any given question. Google AI appears to just make a
guess and pick one -- and about half the time that appears to be
wrong. It's even worse when these threads contain sarcastic responses.
If anything, Google AI Overview answers seem to be becoming less
accurate not more accurate. A true train wreck of AI misinformation
and slop.

L

- - -
--Lauren--
Lauren Weinstein
lau...@vortex.com (https://www.vortex.com/lauren)
Lauren's Blog: https://lauren.vortex.com
Mastodon: https://mastodon.laurenweinstein.org/@lauren
Signal: By request on need to know basis
Founder: Network Neutrality Squad: https://www.nnsquad.org
PRIVACY Forum: https://www.vortex.com/privacy-info
Co-Founder: People For Internet Responsibility
_______________________________________________
nnsquad mailing list
https://lists.nnsquad.org/mailman/listinfo/nnsquad