Has the War Against AI Already Been Lost?

[Unbeliever]

Did you use catsup? 🍊

[Gawdzilla Sama]

Well, that certainly was a question.

[Hydra009]

Catholic AI "priest" defrocked

Artificial intelligence shook up a corner of the Catholic community after Catholic Answers—a San Diego, Calif.-based publisher of church resources—introduced "Father Justin," an AI chatbot on Monday. The virtual priest was defrocked by the end of the week, but its creators will still allow the chatbot to operate as simply "Justin."

Other users shared examples of Father Justin's priestly counsel, including telling one user they can use Gatorade instead of holy water in a baptism and explaining how a brother and sister can prepare for marriage to each other. Father Justin also accepted a user's confession and then told her all of her sins had been absolved.

I mean, is that really so different?  On the plus side, at least there's no rap sheet.

Among the many missteps is just the presentation of what is essentially a chat bot as a sort of digital priest, but without any of the - dare I say it? - soul of a human being.  I don't think anyone would object to a chat bot presented to them as a chat bot.  Humanizing this thing was a big mistake.

In a separate speech in December, Francis called for the ethical development of artificial intelligence to serve humanity and promote peace.

I broadly agree with him there.  Check out the full speech, it's pretty good.

[Unbeliever]

If AIs can be priests, who needs human priests!? 😇

[the_antithesis]

Maybe get robot monks like from Dirk Gently that do all that pesky believing for you.

[the_antithesis]

[Hydra009]

Most people assume the robot uprising will be full of purple laser guns and explosions.  No.  It will be full of glue.

[Cassia]

Most people assume the robot uprising will be full of purple laser guns and explosions.  No.  It will be full of glue.

That is funny.

[Hydra009]

More terrible AI advice from Google

They really need to stop using Reddit to train AI.  :/  There's a lot of insightful and funny content on Reddit, but man oh man, there's a TON of awful advice and negativity.  Going to Reddit for mental health advice is like going to a frat house for alcoholism rehab.

Especially bad advice includes cutting open a snake bite and/or sucking the venom out of the wound (it's physically impossible to de-venom someone that way and it'll almost certainly cause extra harm to them or yourself, especially when you waste time doing that instead of calling the ambulance and identifying the snake species)

And last but not least, I find it especially irksome that google takes such a lackadaisical approach to commonly-asked suicide-related questions, apparently now even offering helpful tips.  :(

"Many of the examples we've seen have been uncommon queries, and we've also seen examples that were doctored or that we couldn't reproduce," it said, adding that AI Overviews underwent "extensive testing" before its launch.

Ah, the classic "that's not happening on my computer" with the implication that these problems are just made up combined with the other excuse that these are just unusual searches that borked the AI that simultaneously was extensively tested before going live.  Lots of mutually-exclusive propositions going on at once.

Either the thing was extensively tested - which would include lots of uncommon queries - or it wasn't.  Either these uncommon queries are producing borked responses or they're not.  Pick a lane.  It'd be like me blaming my bad driving on lack of sleep and bad directions and everyone else being a bad driver while simultaneously also maintaining that I'm a flawless driver who's always aware of the road at all times.

[Hydra009]

I'd also like to point out that all of the awfulness throughout this thread has a single root cause:  the almighty dollar.

A new Google Search AI feature helped boost its first quarter revenues.

That's the sole reason all this stuff is being rushed out without proper vetting or safeguards - profit.  To improve the bottom line and please shareholders.  And if someone eats glue or jumps off a bridge, no one is bothered.

[Mr.Obvious]

One of my favorite shows is called 'De Mol'.
It is a game show in an exotic location and each episode one person has to leave, he or she who answered the least questions right about who 'the mole' is.
Basically, out of the ten people, one is a saboteur who has to screw up as many of the games and missions as possible, preferably whilst being undiscovered for as long as possible so he or she can wreak as much havoc as possible.

This year, the had some clever uses of AI in the missions. They had a mechanical watchdog for a villa they had to infiltrate.
And one of the 'prizes' the contestants could stab eachother in the back for, was to hear the voice of the mole in private. Which the winner of that mission indeed got, but it was flanked by ai verdions of all the remaining voices. So he had a hard time figuring who's wqs real.
Clever trick. Had me giddy.