Chat GPT.

[Milzy]

Who has it? What’s the cheapest package? What if it gets filled with propaganda & misinformation? Is A.I going to be dangerous to humans?

 

[ebikeerwidnes]

I read a few days ago that it works by using the information it can find on the WWW

If you can change that information then the way it works changes based on that

which means any system reliant on AI has to be very careful about the data it uses to makes its decisions

 

[albion]

Yes indeed, AI will certainly be very easily gamed.

However, we are straying into politics.

 

[Drago]

I read a few days ago that it works by using the information it can find on the WWW

If you can change that information then the way it works changes based on that

which means any system reliant on AI has to be very careful about the data it uses to makes its decisions

So it's just like yoomans then.

 

[DCLane]

I've got a small group of students who are using it as part of an official assignment, comparing real-life answers to interview questions to ChatGPT answers.

Oh, and there's students who are using ChatGPT to create their assignments ... without realising there's also a checking mechanism where they submit to identify this.

 

[geocycle]

I've got a small group of students who are using it as part of an official assignment, comparing real-life answers to interview questions to ChatGPT answers.

Oh, and there's students who are using ChatGPT to create their assignments ... without realising there's also a checking mechanism where they submit to identify this.

I see potential to reverse engineer it to do the piles of marking that appear at this time of year!

 

[midlife]

I've got a small group of students who are using it as part of an official assignment, comparing real-life answers to interview questions to ChatGPT answers.

Oh, and there's students who are using ChatGPT to create their assignments ... without realising there's also a checking mechanism where they submit to identify this.

Yep. Chat GPT is a large language model and quite easy to spot if used for essays. Son works as a Dev for Lancaster Uni computer department and they can screen assignments to catch the cheaters.

Might get a bit more difficult if large language models turn into AI but son reckons they will still catch them...

 

[Dogtrousers]

Yep. Chat GPT is a large language model and quite easy to spot if used for essays. Son works as a Dev for Lancaster Uni computer department and they can screen assignments to catch the cheaters.

Might get a bit more difficult if large language models turn into AI but son reckons they will still catch them...

Detecting essays written using AI and sifting them out from human-written ones is an exercise in pattern matching. ... Sounds like perfect job for an AI

 

[ebikeerwidnes]

So it's just like yoomans then.

Exactly - except that it potentially has - and uses - access to the whole WWW
however - if you tell it to use only GB News and the Daily Mail and similar sites then the results will not be the same if the systems uses the BBC and the Guardian etc
(other sites are available!!!)

 

[Pat "5mph"]

Sorry guys, this is really a NACA topic.