SkyNet really is ... here?
- Thread starter Theory of Games
- Start date
Scribe
Legend
Take the echo chamber of social media of 2018-2022, and add a boost of validation to your delusion and mental illness.
www.the-independent.com
And with that, I'm off to the gym. Remember to unplug and engage with reality folks!
ChatGPT adds mental health guardrails after reports of bot feeding people’s delusions
ChatGPT users who converse with the bot for an extended amount of time will now receive reminders encouraging them to take a break
www.the-independent.com
And with that, I'm off to the gym. Remember to unplug and engage with reality folks!
Theory of Games
Storied Gamist
Well. I tried - and I'm done with this thread 
You know, if an AI apocalypse occurs it is most likely to be something odd and unexpected, like in Charlie Stross's Freya 'verse series. Where the humans discovered that AI only really got intelligent in bodies and the original template need a training period similar to educating a human and made sexbots.
The human race dies out due to lack of reproduction.
The human race dies out due to lack of reproduction.
Scribe
Legend
To be clear, that link from yesterday is related to this one.
gnarlygninja
Hero
That seems like something you'd have wanted to include a while ago. Reading the chatlog in that link is revolting.
Scribe
Legend
Yeah by bad, multiple tabs yesterday going over the same story, that new link was just while I was checking if Nvidia had posted their numbers...
gnarlygninja
Hero
Oh, no, I meant I don't know why ChatGPT didn't have anything to stop it from enabling and actively encouraging suicidal ideation.
I don't really understand how a person can read it telling that kid to keep his thoughts a secret, so the space he killed himself would be the first place anyone saw him, and then defend these people. Absolutely disgusting, anti human behavior.
I don't really understand how a person can read it telling that kid to keep his thoughts a secret, so the space he killed himself would be the first place anyone saw him, and then defend these people. Absolutely disgusting, anti human behavior.
well the explanation is easy..... because there are a thousand things that ChatGPT will do if left unchecked that you have to be specifically code to stop, and each one can have a dozen variations that you have to account for.
And that takes time, energy, and money. That in a nutshell is the problem, its a technology moving so quickly that your never going to keep up with all of the safeguards you need to have it operate safely. And that's assuming you even correctly imagine what the safeguards will be, more than likely it will continue to be a string of cases that no one really thought of until it happens. And they will dutiful patch each and every issue....until the issue becomes too grat and something really bad happens.
Similar Threads
