Script from my national radio report yesterday on Google Search AI
Overviews and other AI issues
Here is the script of my national radio report yesterday evening
regarding Google's Search AI Overviews, and other AI issues. As
always there were a few very slight wording differences between this
script and my live presentation of the report.
- - -
So the technical term for what's happening now with AI, especially
generative AI, is NUTS. I mean it's not just Google, but Microsoft
too, with OpenAI's ChatGPT. The firms are just pouring out half-baked
AI systems and trying to basically ram them down our throats whether
we want them or not, by embedding them into everything they can,
including in irresponsible or even potentially hazardous ways. And
it's all in the search of profits at our expense. I'll talk
specifically about Google Search shortly, but so much of this crazy
stuff is being deployed. Microsoft wants to record everything you do
on a PC through an AI system. Both Google and Microsoft want to listen
in on your personal phone calls with AI. YouTube is absolutely flooded
with low quality AI junk videos, making it ever harder to find
accurate, *useful* videos. Google is now pushing their AI "Help me
write" feature which feeds your text into their AI from all over the
place including in many Chrome browser context menus, where in some
cases they've replaced the standard text UNDO command with "Help me
write". And Help me write is so easy to trigger accidentally that you
not only could end up feeding personal or business proprietary
information into the AI, but also to the human AI trainers who Google
notes can also see this kind of data.
OK, now about Google Search. For quite some time now many people have
been noticing a decline in the quality of Google search results -- and
keep in mind that Google does the overwhelmingly vast percentage of
searches by Internet users. So Google has recently been rolling out to
regular Google Search results what they call AI Overviews, and these
are AI-generated answers to what seem like most queries now, that can
push all the actual site links -- the sites from which Google AI
presumably pulled the data to formulate those answers -- actually push
them so far down the page that few users will ever see them, and this
potentially starves those sites that provided that data from getting
the user views they need to stay up and running.
Some of the AI overview answers have links but often they're dim and
obscure and almost impossible to even see unless you have perfect
20/20 vision and very young eyes. On top of that many of these AI
Overview answers are just banal, stupid, and often just confused or
plain wrong, mixing up accurate and inaccurate information, sometimes
in ways that could actually be unsafe, for example when they're wrong
about health-related questions. This is all very different from the
kinds of top of page answers that Google has shown for straightforward
search queries like math questions or definitions of words or when was
a particular film released that they've provided for some time now.
These AI Overview answers are showing up all over the place and like I
said, much of the time their quality is abysmal. Now of course if
you're not knowledgeable about a subject you're asking about, you
might assume a misleading or wrong AI Overview answer is correct, and
since Google has now made it less likely that you'll scroll down the
page to find and visit sites that may have accurate information, it's
a real mess. There are some tricks with Google Search URLs that I've
seen to bypass some of this for now, but Google could disable those at
any time.
What's really needed is a way for users to turn all of this generative
AI content completely off until such a time, if ever, that a given
user decides they want to turn it on again. Or better yet, these AI
features should be *entirely* opt-in, that is, turned off UNTIL you
decide you want to use them in the first place.
So once again we see that fears of super intelligent AIs wiping out
humanity are not what we should be worried about right now. What we
need to be concerned about are the ways that Big Tech AI companies are
hell-bent on forcing generative AI systems into all aspects of our
private lives in ways that are often unwanted, confusing,
irresponsible, or even worse. And the way things seem to be going
right now, there's no indication that these firms are interested in
how we feel about all this. And that's not going to change so long as
we're willing to continue using their products without making it clear
to them that we won't indefinitely tolerate their push to stuff
generative AI systems into our lives whether we want them there or
not.
- - -
L
- - -
--Lauren--
Lauren Weinstein
[email protected] (https://www.vortex.com/lauren)
Lauren's Blog: https://lauren.vortex.com
Mastodon: https://mastodon.laurenweinstein.org/@lauren
Founder: Network Neutrality Squad: https://www.nnsquad.org
PRIVACY Forum: https://www.vortex.com/privacy-info
Co-Founder: People For Internet Responsibility
Tel: +1 (818) 225-2800
_______________________________________________
google-issues mailing list
https://lists.vortex.com/mailman/listinfo/google-issues
