I can tell you that most of the level of management who are mandating this sort of thing aren’t thinking it through that far. “They have it, we have to have it” is as far as they can comprehend it.
If they had the foggiest idea what it is, how it works, or what it’s capable of (both in the positive and negative senses), it wouldn’t be the thing that it is, but they don’t, and it is, and everybody who does have an idea feels like we’re all taking crazy pills while everyone goes mad around us.
The people who are selling it to them, they’re working on the basis that the more people interact with the machine, the faster the machine learns. Because now that they’ve sucked up the entire internet, their new training resource is in the chat logs from the last iteration of the engine.
You mean everyone has access to their own personal conversational partner who is confident, authoritative, utterly without shame or embarrassment, and completely untethered from reality?
I’m concerned about the folks who want the data for other reasons. Sometimes hype about creating artificial intelligence is just cheese in the data-grabbing trap. This video about a “hidden” Ancestry site feature had me smh:
I avoided the company’s last pitch about getting older relatives to recite and save family stories on their site, because of the potential alternative uses for voice recordings. I’m not into posting people’s pictures, either. Gotta find out what happened to the lawsuits filed after Ancestry started scooping up and publishing people’s high school yearbook photos. I was a minor when mine was taken, and did not consent to posting it online.
Although that site is supposed to focus on information about people who are deceased, the number of records for the living (like voter registration) is increasing at an alarming rate. I tried the feature described in the video on an image someone else posted of a cousin who passed away 30 years ago. A little message pop-up saying you agree this is not supposed to be done with living people is all that needs to be clicked before you see potential matches. Just like on social media sites, privacy goes out the window if you happen to be in a photo taken and posted by someone else. AI + facial recognition + DNA matches* =
*Not mine, I manage records from my father. I don’t add information about anyone living who is younger than his generation, either.
As always with “AI” when its output is what would be expected (stochastic parrot) we say “isn’t it great it can replicate human stuff?” When it outputs nonsense we parse it for it and say “isn’t it creative?”
But in a recently published article on “The Law of Digital Resurrection,” law professor Victoria Haneman warned that “there is no legal or regulatory landscape against which to estate plan to protect those who would avoid digital resurrection, and few privacy rights for the deceased. This is an intersection of death, technology, and privacy law that has remained relatively ignored until recently.”
Yeah, no, please…
She told Ars that right now, “a lot of people are like, ‘Why do I care if somebody resurrects me after I’m dead?’ You know, ‘They can do what they want.’ And they think that, until they find a family member who’s been resurrected by a creepy ex-boyfriend or their dead grandmother’s resurrected, and then it becomes a different story.”
For the fun of it, I tried it out… with the Titanic swimming pool question. The fake podcasters disagreed with the AI overview: “Well, the literal answer is yes, it’s full of water. I mean, the entire wreck is submerged, so the pool is definitely not dry.” And then it went into a bunch of other stuff about how first class passengers had to pay to swim in it, and that swimsuits were provided, and one of the survivors mentioning taking a swim the day before the sinking (“wow, that’s eerie”, “you can say that again”… uh, no idea why someone mentioning they took a swim the day before the sinking is at all eerie, but ok…)
I do not see why the heck I would want to spend several minutes listening to voices chat back and forth with each other, complete with jokes and irrelevant asides, when I could just be looking at the actual search results. Even if I was in a situation where I couldn’t read the results, this would just waste too much of my time to be at all useful. And that’s before noting that it took two tries (the first one died with an error) and at least a whole minute before I could even listen to that.
[edit] After typing that transcription above, I had that “did I actually spell that right” moment with the word “eerie”… so I searched for just that word. And then, for the heck of it, I tried the podcast result. 3.5 minute long podcast just chatting about the definition of the word “eerie”.
“It’s more than just being scared. It’s a specific kind of unsettling feeling, a sense that something is off, that the world isn’t quite behaving as it should. It’s like a whisper of the uncanny” Well, it hit that nail on the head, without even knowing it. And, even stranger… the male podcaster’s voice abruptly changed partway through the recording, suddenly becoming a very slightly higher pitch and more nasally for several sentences of back-and-forth, before switching back, all while talking about how things can feel odd and disconnected. Felt like a third, completely separate person had suddenly jumped into the conversation with no acknowledgement, and just as suddenly left again.
Gotta find out what happened to the lawsuits filed after Ancestry started scooping up and publishing people’s high school yearbook photos. I was a minor when mine was taken, and did not consent to posting it online.
Just like on social media sites, privacy goes out the window if you happen to be in a photo taken and posted by someone else. AI + facial recognition + DNA matches* = 
