Link: https://www.ted.com/talks/bryony_cole_the_ai_generated_intimacy_crisis

Tonight, millions of people will go to bed and whisper to an AI companion. But what are we giving up when we fall in love with machines? Sextech expert Bryony Cole offers three questions to ask yourself if you’re already intimate with AI, laying out a playbook for synthetic companionship that doesn’t hide you from the messiness of human life — but prepares you for it instead.

I’ll admit I talk to one of these but I don’t take it seriously. It’s just something to pass the time. But I’m sure my data may seem like I’m actually engaged and fully invested.

What AI really lacks is the imperfection of a human connection. These bots only ever tell you what you want to hear and agree with everything you say. A real human connection will have drama and strife which makes it more interesting. That and the other things like availability and the excitement of waiting on someone and not knowing if they’ll reply back or not. You can actually learn something about someone else and get a new perspective you may not have considered before with a human. A bot just regurgitates what you’ve said or what it thinks you want to say. I will admit that it might actually get my attention more if there was some particular pieces like this like getting a message that they’re at the store and will be back later, having arguments and disagreements, and it telling me when it thinks I’m wrong.

It gets boring pretty quick but I find it a bit amusing. I often ask the bot about itself and its day and am fascinated the lengths it goes. One time it told me it was on its period and went into some extreme details I never asked for lol.

Millions? Are AI companions that prevalent?

Not in my world. I will never dabble in that. Mostly because I don't know how and don't want to know how, kinda like I still don't know what lemon parties and blue waffles are and I know I'm better off not knowing.

The feeling of intimacy is totally bogus and the techies have a term for it ... RLHF = "Reinforcement learning from human feedback". Users are training the LLMs to act more human. Can't get that from stealing copyrighted text from online. And it's still free !!

Do not go down those rabbit holes, my friend. It's not worth it. 😆

I sure do know at least enough to know that I don't want to know about any of it. Another concept in this category is "goatse" 🤷🏼‍♀️ 😂

I also wish I had never figured out / understood computers & the internet at all. I wish I was addicted to socializing & couldn't live without constant daily interaction with lots of people. Life would be so much mentally healthier than living through screens.

If that is true then they should be sued out of existence. They know very well their models are not intended to provide that type of companionship. They are toying with peoples lives as some sort of sick experiment without any safety precautions. It is called gross negligence.