Script of my radio report regarding digital/virtual assistants (and Google)
This is the script of my national network radio report yesterday
discussing (unconfirmed) reports that Google is working on an
AI-based digital/virtual assistant, and discussion regarding these
assistants more broadly. As always, there may have been minor wording
variations from this script as I presented this report live on air.
- - -
So this is one of those situations where you think you've already
heard all the awful and risky applications of AI, and then along comes
another one that's even worse. And this one is something that we knew
was coming sooner or later, unfortunately it seems to be sooner rather
than later.
And what we're talking about right now falls under general category of
AI-based so-called "digital assistants" also called "virtual
assistants". And I want to be clear that these reports that Google is
working on this and may even be quite close to an announcement have
not been confirmed by Google at this point as far as I know. But it is
logical that they'd be working in this area and we know that other
AI-companies are working along these lines.
So yeah, the idea is to bring to reality the Big Tech AI fantasy of a
digital, virtual "assistant" or agent that can take over much of your
life, at least to the extent that your life is trapped in your web
browser. The general concept is that you let these AI systems read
everything on your screen so they know everything you're doing, and
then you can ask them to do stuff via the browser for you that you'd
otherwise do yourself. So a kinda classic example would be telling the
AI you want to visit New York City from L.A., and then the AI looks up
ticket prices and hotel prices and then goes ahead actually buys the
tickets in your name and books the hotel room for you and all you have
to do is smile.
Now of course a whole list of potential problems comes up even beyond
the obvious privacy issues associated with letting an AI have that
kind of access to what you're doing on your browser. And obviously if
the AI is going to be conducting financial transactions for you, any
misunderstandings or mistakes it makes could have pretty serious
negative implications ... for you .. not for the AI.
Because it seems rather unlikely that any kind of serious
responsibility would be taken by these firms for errors that their AIs
make along these lines, at least judging from their current attitudes
about errors and misinformation copiously flowing forth from their
chatbots and such.
Personally, even if these AIs were less error prone, I wouldn't want
an AI digital assistant. Because you're the one who's going to end up
ultimately having to deal with whatever decisions the AI makes that
turn out not to have been in alignment, shall we say, with what you
actually wanted to do.
And all this isn't even taking into account the ramifications if
hackers got into these AI assistant dialogues and action sets, for
example pretending to be you and ordering the AI to do any number of
things in your name that wouldn't be to your benefit.
Here's an example of something about AI mistakes. Over the weekend I
tested Google Gemini with an easy fact-based question about the
classic 1956 sci-fi film "Forbidden Planet". And I asked it if Dr.
Morbius was the only human on the planet when the relief ship
arrived -- the correct answer is no, his daughter Altaira was there too.
But Gemini replied that he was alone. So I asked what about his
daughter? Wasn't she there also? And Gemini replies, oh yeah, actually
she was and is a main character in the film. And I asked, so why did
you claim Morbius was alone just now? And Gemini says, sorry for the
oversight.
Gemini may be sorry but I don't trust these answers even to simple
fact questions because these systems are often just plain wrong, and I
certainly would never trust them about anything important.
So do you really want these systems acting as your digital assistant
controlling your web browser? One thing's for sure, I don't want
ANYTHING to do with them -- period!
- - -
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
_______________________________________________
pfir mailing list
https://lists.pfir.org/mailman/listinfo/pfir
