AI - is it just, you know, bollocks?

[rick49]

"For example, someone on Twitter posted an example of asking GPT-4 to write instructions for how to hack a computer, and it provided some terrifying detail."
www.digitaltrends.com/computing/is-chatgpt-safe/

doesn't sound like a harmless toy.

[rick49]

www.techradar.com/news/chatgpt-is-now-being-used-to-make-scams-much-more-dangerous

could render our ability to determine whats real and whats not real useless.

[perrykneeham]

FFS stop being frightened of everything Rick.

[flatandy]

May 20, 2023 17:17:32 GMT rick49 said:

"For example, someone on Twitter posted an example of asking GPT-4 to write instructions for how to hack a computer, and it provided some terrifying detail."
www.digitaltrends.com/computing/is-chatgpt-safe/

doesn't sound like a harmless toy.

Presumably because the info is already out there. The only way it can know is if it's repeating what was already findable via search engines.

[rick49]

May 20, 2023 17:37:28 GMT perrykneeham said:

FFS stop being frightened of everything Rick.

when the creators themselves are afraid of their creation and warning us about it, we should be worried. not that their fears and warnings will be listened to, because they won't be. it's already too late.

[perrykneeham]

Pfft! Are they actually afraid?

[rick49]

yes. i posted the links.

[perrykneeham]

You know how people who say they're God-fearin' aren't actually afearin' of God?

[rick49]

May 20, 2023 19:28:49 GMT perrykneeham said:

You know how people who say they're God-fearin' aren't actually afearin' of God?

not familiar with that one. lol but when ai reaches the point where it can learn and teach itself, it's game over. we will be given the pink slip.

[perrykneeham]

Well, I won't be losing any sleep over it.

[voice]

What ever is called AI presently is simply not, it's certainly not intelligent, it can't fathom, reason or critically think. It's a poor reflection of the crap floating around the internet.

Case in point, I heard they tried a set up am AI learning thing on social media, set this thing off to see where it went. They had to turn it off after only a couple of days as it quickly glommed onto a married of conspiracy theories, became incredibly anti Semitic and racist as the social media algorithms fed it an increasingly toxic mix.

If what we were seeing was an actual intelligence there might be concerns about them taking over, but being frightened by chatboxGBT or anything similar is stupid. The real dangers appear to be students using it to cheat and never winning at chess. It's nit going to put any artist or writer of note out of a job.

[rick49]

so the creators of ai issuing warnings are full of crap and don't know what they are talking about.

[voice]

What they want regulating is not the same stuff you appear terrified of.

Mainly cos they know it's not intelligent in any conceivable way, but is a blunt tool that can produce huge quantities of crap and disinformation, because basically it reflects the dross on the internet, the good the bad and the ugly. It fools simpltons but anyone with a modicum of nouse really can tell something created by chatbotGPT when compared to something writer by a person.

[perrykneeham]

I think there is an element of trade protectionism at work too: not because AI is likely to take work away from certain trades, but because they may be required to do their own jobs better.

The teachers are shitting themselves but that's not because they're worried that the students' will cheat. Actual physical exams will ultimately deal with that and there is already all sorts of plagiarism software.

[ootlg]

In China they have a national security system which monitors the movements and activities of every Chinese person registered in the system(ie. all but the illegals, ironically). It runs itself and keeps an eye on everything. Now here's the thing: when the computer's algorithms detect unusual activity from someone, it automatically issues an arrest warrant on that person.

Coming our way sooner or later?

[perrykneeham]

I'm not sure I have a problem with that. It's all about the criteria you choose and oversight.

[ootlg]

It may indicate the potential for problems with AI... misspelt name, incorrect programming and you could find yourself being held albeit innocently for something you didn't do. Loss of job, position in society, inability to open a bank account etc etc. You could change your car and the changeover dates might not coincide quickly enough and you're suddenly wanted for armed robbery.

[flatandy]

All those problems exist without AI. It's not like Kafka was aware of the dangers of poorly programmed databases and piss-poor AI.

[ootlg]

How could he be?

[flatandy]

My point being that Kafka's "The Trial" pretty much tells the story you're describing 3 posts ago.