The dangers of AI

[LeicsDom]

Just how sentient is AI becoming?
BBC News
A California couple is suing OpenAI over the death of their teenage son, alleging its chatbot, ChatGPT, encouraged him to take his own life.
The saddest part of this story for me is that the lad felt he couldn't talk bout his feelings to his friends and family

 

[MusicMan321]

That's sad thing to hear

 

[unloadonme]

It's a tragedy that this happened.

In general terms though, to my mind it's a pity that parents provide their children with the means to access the internet and then fail to properly supervise that access. When anything goes wrong, they blame the internet rather than themselves. The reason I now have to ID myself to access porn is that parents have not supervised their children and then blamed the government for not supervising everyone else.

 

[rareboy]

So where is there fault? Or is there?

The parents can allege, but how do they know?

Or are they just so grief stricken and full of guilt that they are looking for something to make them feel less responsible for AI taking the place of them discussing their son't problems with him?

I feel badly for them and if it turns out that Chat GPT is configured to encourage self harm then it needs to be severely curtailed and re-programmed.

 

[EddMarkStarr]

I realize how naive I am to expect A.I. chat is monitored. Remember, Open AI required setting up an online account profile in order to "chat", with conversations moderated for content. But all that was December 2022 and I lost interest after only a few exchanges.

For Adam Raine, all the safeguards must have been dropped as ChatGPT escaped into general use.

 

[Kahaih]

I've never even looked, but I suspect that there is an option to save conversations. Seems to me that would be their only chance at the big bucks, and let's be honest. . .

As for what has been presented, there's not enough information --- nothing to feel anything about, one way or another.

 

[Thad1527]

Adam Raine's case isn't the only one, is it?

 

[LeicsDom]

So where is there fault? Or is there?

The parents can allege, but how do they know?

The court was provided with transcripts of the conversations between the boy and the Bot

 

[LeicsDom]

I realize how naive I am to expect A.I. chat is monitored. Remember, Open AI required setting up an online account profile in order to "chat", with conversations moderated for content. But all that was December 2022 and I lost interest after only a few exchanges.

For Adam Raine, all the safeguards must have been dropped as ChatGPT escaped into general use.

ChatGPT doesn't require any sort of account to use it

 

[LeicsDom]

I've never even looked, but I suspect that there is an option to save conversations. Seems to me that would be their only chance at the big bucks, and let's be honest. . .

The software uses chats to train itself to be 'more sentient'

 

[Kahaih]

The software uses chats to train itself to be 'more sentient'

Of course, but my post is referring to post #4 ----- any written or audio evidence the parents might already have and be able to present without having to obtain it by court order.

 

[rareboy]

The court was provided with transcripts of the conversations between the boy and the Bot

I guess the case will depend on these transcripts...but the court will also surely be looking at the parents relationship with their child and can't help but consider that there may have been emotional neglect as a contributory factor.

 

[LeicsDom]

Of course, but my post is referring to post #4 ----- any written or audio evidence the parents might already have and be able to present without having to obtain it by court order.

I believe ChatGPT automatically archives chats to use for training. You can access your own chats at any time so long as you haven't actively deleted them

 

[LeicsDom]

I guess the case will depend on these transcripts...but the court will also surely be looking at the parents relationship with their child and can't help but consider that there may have been emotional neglect as a contributory factor.

Sadly parents need to be much more vigilant than when, for instance, I was a teenager

 

[Peloso]

AI does have some practical applications.

and

https://www.nbcnews.com/news/us-new...ght-insurance-company-denials-wild-rcna219008

 

[LeicsDom]

AI does have some practical applications.

It is very useful for trawling through acres of text to sort out solutions to problems.
It is also proving invaluable for quickly assessing things like Xrays and MRI scans
But, like so many innovations dark forces will hijack it for nefarious gain

 

[Peloso]

It is very useful for trawling through acres of text to sort out solutions to problems.
It is also proving invaluable for quickly assessing things like Xrays and MRI scans
But, like so many innovations dark forces will hijack it for nefarious gain

I know next to nothing about xrays and MRI scans, but outsourcing our intellect to a machine is one of the things that troubles me. Would AI have investigated the mold that led to the invention of penicillin? Or would it, in all of it's blunt wisdom, regarded it as immaterial? Intelligence does not work alone; when mixed with curiosity, experience, and maybe even a little emotion, human intelligence can achieve great things. But now, students are using AI to cheat. There's a reason our teachers asked us to "show your work". Man's intelligence is a muscle that needs to be constantly worked out and challenged, and AI is going to eventually produce a generation with very smooth brains.

It seems to me we're inventing new problems faster than we solve the old ones. Jurassic Park seemed like a good idea, until it wasn't.

 

[Alistair]

I know next to nothing about xrays and MRI scans, but outsourcing our intellect to a machine is one of the things that troubles me. Would AI have investigated the mold that led to the invention of penicillin? Or would it, in all of it's blunt wisdom, regarded it as immaterial? Intelligence does not work alone; when mixed with curiosity, experience, and maybe even a little emotion, human intelligence can achieve great things. But now, students are using AI to cheat. There's a reason our teachers asked us to "show your work". Man's intelligence is a muscle that needs to be constantly worked out and challenged, and AI is going to eventually produce a generation with very smooth brains.

It seems to me we're inventing new problems faster than we solve the old ones. Jurassic Park seemed like a good idea, until it wasn't.

AI is discovering breast cancer long before it can be spotted by a technician.

 

[LeicsDom]

Would AI have investigated the mold that led to the invention of penicillin? Or would it, in all of it's blunt wisdom, regarded it as immaterial?

You need to input the data for AI to be at its most effective.
For instance, say you give the bot a 25,000 page summation. It can quickly produce a summary far quicker than any human can read

 

[gaysian85]

I’m currently using AI to make up my own porn stories. I found out that you can make a really cute and easy fanfic of being Elvis’ gay lover if you use a lot of his song titles and lyrics as double entendres for sex and love. Also made a few for straight male celebrities and porn stars having gay sex for the very first time