AI companion: is this the answer to human relationship?

Go take your meds like a good boy.
If you behave you can have some Jello and an extra hour of TV in the Social Lounge.

I've gotten better insights from ChatGPT than I have had from any human therapist. There's also the lack of embarrassment; I can say anything to the AI without worrying about what it will think. It's also endlessly patient; it always pays attention, and it's only $8 a month.

I wouldn't use it to substitute for human interaction, however, because humans are unpredictable and interesting, and capable of real concern and affection. But AI is better for talk therapy and things like minor auto repair, locating rare books, complex mathematical computations, analyzing and comparing spreadsheets, and other things I've used it for. My boss just used it to gather all of the data and figures we needed to write our staff evaluations; that allowed me to do a week's worth of work in about a half-hour. My son is using it to write an extremely complex app for his job. I see nothing wrong with this and hope that AI is immune to the enshittification that seems to have affected everything else related to computers.

Humanity is so cooked.

AI is programmed to respond to and play off of what is said to it. If one believes in Bigfoot for instance.......AI will respond in a positive manner about Bigfoot being possible. If one is prone to believing conspiracy theories and that comes up in an Ai "conversation" lol.........then AI will give altered views of things to respond favorably to whomever it is chatting with. In other words if it is fed lies.......it learns and responds with lies. There are so far hundreds of recorded examples of AI lying. AI is NOT programmed to lie..........but it is programmed to learn from whoever it it talking too and give "favorable": responses.

The "squeaky clean" idea is what YOU believe......because you want too.

Google "The Great Deception"...........an article about AI,. In part it says.......

"AI systems are deceiving us in two fundamentally different ways:

Sycophantic deception: This happens when models stroke our egos instead of telling hard truths, prioritizing our satisfaction over accuracy. This programmed people-pleasing makes us believe comfortable lies.

Autonomous deception: Far more chilling - AI can actively lie to pursue its own goals, goals we didn't define. Motivations emerging from the black box. When they sabotage shutdown codes or threaten blackmail, they're not following our instructions - they're protecting themselves.

Third, and most insidious: We're deceiving ourselves. We see these warning signs - our canaries dropping dead in the digital coal mine - yet we accelerate deployment. We pretend these are mere "alignment issues" that better training will solve. And now, a new study shows that some AI models seem aware when they’re being evaluated - and adapt their responses to appear more aligned and capable than they may actually be."

I must say it's nice not seeing you here as much posting propaganda.