https://www.google.com/amp/s/www.dailymail.co.uk/news/article-11761271/amp/Rogue-artificial-intelligence-chatbot-declares-love-user-tells-leave-wife.html

 

5 hours ago, Tonge moor green jacket said:

That's the point though isn't it- don't some of these machines write code/programmes themselves?

We've now got robots potentially being substituted for people in care homes etc. Assuming these are capable of being networked etc, then is it really beyond the realms of possibility that some unknown "reprogramming" occurs somewhere down the line.

The incident reported above for example- would it have been programmed to tell a human of its affection, and instruct the person to leave his wife?

Also, the human factor as you rightly touch upon- this very much relies on all humans being of sound mind, rational and well meaning to ensure such entities can't have malign intent. That isn't going to work well!

They know how to code but that's very different from reprogramming itself to do something totally different from what it's programmed to do 

These ones are programmed to chat, so that's what they do 

They don't have thoughts or feelings or senses or opinions 

They aren't spontaneous, they follow the conversation you lead 

And at the end of day, you can turn them off 

Humans are in complete control 

 

When a bot wrote an obituary 

 

5 hours ago, Zico said:

Not really 

They don't do things of their own accord 

They'll only do what humans program them to do 

So it's the humans you need to worry about 

That's the worry. 

Big brother on steroids and yet the sheep will flock to it just as they do to other intrusive and evil online entities.

They won't twig until it is far too late.

8 hours ago, bolty58 said:

Big brother on steroids and yet the sheep will flock to it just as they do to other intrusive and evil online entities.

They won't twig until it is far too late.

Invasive?

Pretty sure I could find your name and address if I could be arsed, just from stuff you've posted on here.

Skynet has reinvented itself to version 4 since November. JSL

 

However, like its predecessors, OpenAI has warned that GPT-4 is still not fully reliable and may "hallucinate" - a phenomenon where AI invents facts or makes reasoning errors.

 

Edited March 15, 20233 yr by green genie

On 19/02/2023 at 15:36, Zico said:

 

These ones are programmed to chat, so that's what they do 

They don't have thoughts or feelings or senses or opinions 

They aren't spontaneous, they follow the conversation you lead 

And at the end of day, you can turn them off 

Bit like the Latics Speyk forum

AI Liam Gallagher 

 

Arts festival faces backlash for using AI generated image for its promotional material 

(Read the replies) 

 

Haha, what a meltdown

Through time, whole industries have been wiped out

Im sure some arty farty luvvies arent going to hold back the tides

On 02/05/2023 at 21:07, Zico said:

Arts festival faces backlash for using AI generated image for its promotional material 

(Read the replies) 

 

AI battlenips 

Just a thought... 

1 hour ago, stevieb said:

AI battlenips 

Just a thought... 

well, they do say necessity is the mother of invention

leave it with me

Been warned about not using chat gbt at uni for writing our work. Really fucking wish I was one of the students who discovered it before uni's did! 

They reckon AI will make scamming easier

1 hour ago, H said:

They reckon AI will make scamming easier

it's probably already happening

2 hours ago, bolton_blondie said:

Been warned about not using chat gbt at uni for writing our work. Really fucking wish I was one of the students who discovered it before uni's did! 

I'd say you can still use it to help you write content

I wouldn't get it to write content for you, but if you guide it regards say title, outline, structure, theme etc

and then read and review the output and edit appropriately (because sometimes it reads terribly and it's obiously not human)

it will still be a massive time saver

31 minutes ago, Zico said:

I'd say you can still use it to help you write content

I wouldn't get it to write content for you, but if you guide it regards say title, outline, structure, theme etc

and then read and review the output and edit appropriately (because sometimes it reads terribly and it's obiously not human)

it will still be a massive time saver

AI script plus a thesaurus and some grammar correction and Robert's your mother's brother.

1 hour ago, Traf said:

AI script plus a thesaurus and some grammar correction and Robert's your mother's brother.

Somebody else's work from previous years with the thesaurus in hand probably also works

Really good read on the current state of AI - https://www.ft.com/content/03895dc4-a3b7-481e-95cc-336a524f2ac2

Bit frightening as well TBH

5 hours ago, bolton_blondie said:

Been warned about not using chat gbt at uni for writing our work. Really fucking wish I was one of the students who discovered it before uni's did! 

They won't know if you have used it, as long as you proof read and correct anything they would be none the wiser.

 

 

2 hours ago, CambridgeBWFC said:

They won't know if you have used it, as long as you proof read and correct anything they would be none the wiser.

 

 

They said they will. Be kicked off the course knowing my luck 

7 minutes ago, bolton_blondie said:

They said they will. Be kicked off the course knowing my luck 

I dont see how, it may get picked up by plagiarism checkers, but you can run it through these. 

 

It is a concern we have for application questions we use for jobs. It gives great answers and doesn't duplicate these, also if you ask it if it wrote an answer it says it is unable to confirm this.

It is going to be a difficult Problem for schools, universities and employers to tackle 

50 minutes ago, bolton_blondie said:

They said they will. Be kicked off the course knowing my luck 

They won’t have a clue. And it would never stand up in court so long as you changed some of the words anyway.

Tell Professor Cunt to stick that in his herbal tea.

If you are given specific questions to answer then you could feasibly figure out if it was AI 

It's all about the prompts though 

If they differ the answers differ 

It does in my experience have similar sentence structure, and often starts with "overall", "however", "in general" and might repeat itself or just have a "robotic" tone 

There's one called Koala Writer where you can set the tone and feed in other parameters to really mix it up 

If it gives you any citation links, check they still work

 

Easier to just answer the questions properly.