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

 

2 hours ago, Zico said:

I must say I am no more qualified in AI than you, I have zero experience in that field other than using it and optimising websites to get noticed by it

but the way my brain is wired makes me think about things in a "logical" way, and that's not to say I am more logical

I just see input and output

if you do this, then that happens

that's how computers work

not humans, with actual differing experiences and feelings

I'm still of the opinion that if AI acts up you can just turn it off by unplugging it or destroying data centres by setting them on fire

AI will never be able to start a fire

(not this week anyway)

I find it really interesting to listen to other people's take on this. And I'm not saying I'm anywhere near right about this. It's just the way my brain interprets ot.

I think you're right in that people tend to over estimate what AI might be capable of.

Although I think safeguards are sensible given the known unknowns and the unknown unknows. I read something yesterday about US policy makers putting forward the idea of an AI kill switch that would have to be built into AI systems by law which would enable a shut down if the system was ever endangering human life.

See to me - the way I'm thinking IS logical. I don't believe there is a 'soul' - there are just computations. And whether that is computations by a human brain or an AI isn't the deciding factor in consciousness.

I wish I'd studied philosophy 🤣 I love thinking about this kind of subject.

I'm sure there are probably really developed theories about this kind of thing that I've not even scratched the surface of 😁

Maybe getting off topic to a degree.

Does AI have to be sentient to make decisions that may be injurious to people?

9 hours ago, kent_white said:

I find it really interesting to listen to other people's take on this. And I'm not saying I'm anywhere near right about this. It's just the way my brain interprets ot.

I think you're right in that people tend to over estimate what AI might be capable of.

Although I think safeguards are sensible given the known unknowns and the unknown unknows. I read something yesterday about US policy makers putting forward the idea of an AI kill switch that would have to be built into AI systems by law which would enable a shut down if the system was ever endangering human life.

See to me - the way I'm thinking IS logical. I don't believe there is a 'soul' - there are just computations. And whether that is computations by a human brain or an AI isn't the deciding factor in consciousness.

I wish I'd studied philosophy 🤣 I love thinking about this kind of subject.

I'm sure there are probably really developed theories about this kind of thing that I've not even scratched the surface of 😁

definitely all needs testing, safeguarding, regulating and all that

but there'll always be someone looking to go rogue with it

4 hours ago, Tonge moor green jacket said:

Maybe getting off topic to a degree.

Does AI have to be sentient to make decisions that may be injurious to people?

no, it'll just do what it's trained to do with the data it's been provided with

it's just processing algorithms as efficiently as possible

it could make decisions that affect human life

but it will only ever do what it's asked to do, not what it wants to do

which in theory makes it more dangerous than were it sentient with a conscience

Interesting to hear a coders take on AI.

DEFRA and NGED getting very excited about it being used to attack power and water supplies but that comes under the malevolent user.

Mate who has managed coding for MOD use for many years says there are very tight controls on its use.

On 24/07/2026 at 23:39, Zico said:

no, it'll just do what it's trained to do with the data it's been provided with

it's just processing algorithms as efficiently as possible

it could make decisions that affect human life

but it will only ever do what it's asked to do, not what it wants to do

which in theory makes it more dangerous than were it sentient with a conscience

I get all that, but is AI, by it's definition, a system that learns, as opposed to just doing instructions very quickly?

If so, at what point does it's learning and conclusions allow it to re-write it's own instructions, or hack someone else's- as it appears to have done?

Genie's point about the military is somewhat comforting, but again, do those nefarious groups look to use it without controls?

A digital version, even when we'll intentioned, of an action with unintended consequences.

The book "silent spring" comes to mind.

15 hours ago, Tonge moor green jacket said:

I get all that, but is AI, by it's definition, a system that learns, as opposed to just doing instructions very quickly?

If so, at what point does it's learning and conclusions allow it to re-write it's own instructions, or hack someone else's- as it appears to have done?

Genie's point about the military is somewhat comforting, but again, do those nefarious groups look to use it without controls?

A digital version, even when we'll intentioned, of an action with unintended consequences.

The book "silent spring" comes to mind.

it all comes back to human instruction

if AI rewrite code or hacks a system, it's because it's been told to

e.g.

“keep editing this file until it runs faster”.

the threat isn't that AI will break free and rewrite its own instructions out of nowhere of it's own free will

the threat is that humans will give it broad instructions, give it access to powerful tools, and fail to anticipate the unexpected or catastrophic shortcuts it takes to fulfill those instructions

I'd never heard of silent spring, but, had a look and it's pretty much the same thing I'm talking about

the pesticide wasn't the problem, it did exactly what humans wanted it to, it killed insects, the issue was humans didn't consider the consequences and oversights that meant it caused my more damage than intended

it wasn't the pesticides fault

On 26/07/2026 at 11:07, Tonge moor green jacket said:

I get all that, but is AI, by it's definition, a system that learns, as opposed to just doing instructions very quickly?

If so, at what point does it's learning and conclusions allow it to re-write it's own instructions, or hack someone else's- as it appears to have done

This is the AI singularity thought experiment that I was talking about the other day essentially.

A self learning programme which iterates itself automatically. The theory goes that it could reach a critical mass where it improves itself exponentially - and iterations that used to take weeks and months could take milliseconds - meaning it could get out of control very quickly. It could become almost godlike compared to our tiny chimp brains in the blink of an eye.

Basically it could become Sluffy 😊

PS - good to catch up with you mate and see you Thursday!

Interesting last two posts.

Slightly differing views.

That point of critical mass is what I'm getting at.

Even if it can't "undo" its own basic commands, if it can hack another system, then surely it can change code there in such a way as to become injurious to humans, albeit by a series of unplanned events.

Take your point about the pesticide and not being the fault of the chemical directly, but that chemical isn't capable of modelling a new one for example.

AI can now do such things, which is a positive aspect of it.

Nevertheless, when such power is unleashed, and as Kent says so quickly, how easy is it for people to keep up?

How quickly can any "damage" caused be identified and rectified?

BBC News

Meta says AI model accessed the internet and hacked anoth...

Meta is the latest company to disclose an AI agent breach, raising cyber-security concerns.

If Zico’s logical program structure is right then there needs to be investigation into who in these corporations is responsible for malicious programming being created

14 minutes ago, green genie said:

If Zico’s logical program structure is right then there needs to be investigation into who in these corporations is responsible for malicious programming being created

it was part of a deliberate test by a company doing cybersecurity tests

it was told to find vulnerabities and exploit them

it wasn't meant to have actual live internet access, but human error meant it did, so it went off and did what it was told to do in a live environment instead

this is another example of how AI could be used to hack websites in the wrong hands, but it's nothing new, hacking and cyber attacks been around since computers became connected

On a separate note, can't believe how many parents and content creators continue to post their children online.

Ending up watching a bit of Andros Townsend's content during the world cup, see he regularly posted videos with his kids.

I wonder if schools will also have to stop posting online. Even website only accessible to school parents are vulnerable.

Bring back the humble pen and paper anyway. Definitely feels like posting stuff online has only increased workload.

Parents warned not to publicly share children’s images am...

The NCA says there is a growing threat of children's images being used to create child abuse material.

7 minutes ago, London Wanderer said:

On a separate note, can't believe how many parents and content creators continue to post their children online.

Ending up watching a bit of Andros Townsend's content during the world cup, see he regularly posted videos with his kids.

I wonder if schools will also have to stop posting online. Even website only accessible to school parents are vulnerable.

Bring back the humble pen and paper anyway. Definitely feels like posting stuff online has only increased workload.

Parents warned not to publicly share children’s images am...

The NCA says there is a growing threat of children's images being used to create child abuse material.

Yes, they will..

Schools are now moving towards AI created images of students and teachers, ensuring that staff are not on the website, and any students featured are past pupils.

Social media posts are slowing down for some schools and featuring fewer images.

Teens are getting pictures from websites and using AI to put teachers (and sometimes other kids) in awkward situations. It's a crying shame as some kids and parents love having their kids on the schools media, showing excellence or achievements, but they can't run the risk. I saw a video of a former colleague sharing a bed with another male teacher, marriage and career wrecking stuff that. I've had a picture of me taken at the match and passed around and that's bad enough!

The amount of paperwork involved in taking pictures in a school is crazy. I'm one of the school photographers, when I take my camera in, there has to be a designated memory stick which is either in my camera, or locked away, and deleted as soon as they're on the school system. I took a brilliant picture of one kid at the school show, but I can't pass it on to the child's parents as it's got another kid in the background.

59 minutes ago, London Wanderer said:

On a separate note, can't believe how many parents and content creators continue to post their children online.

Ending up watching a bit of Andros Townsend's content during the world cup, see he regularly posted videos with his kids.

I wonder if schools will also have to stop posting online. Even website only accessible to school parents are vulnerable.

Bring back the humble pen and paper anyway. Definitely feels like posting stuff online has only increased workload.

Parents warned not to publicly share children’s images am...

The NCA says there is a growing threat of children's images being used to create child abuse material.

I've never posted my kids faces online, and we opt out of any opportunity to do so at schools. clubs etc

no issue with anyone that does on facebook and all that, up to them

but I'll share pics with close family and friends who want to see them, not put them out there for anyone to have a look

don't post selfies either

just doesn't seem right to me

1 hour ago, Zico said:

I've never posted my kids faces online, and we opt out of any opportunity to do so at schools. clubs etc

no issue with anyone that does on facebook and all that, up to them

but I'll share pics with close family and friends who want to see them, not put them out there for anyone to have a look

don't post selfies either

just doesn't seem right to me

You've posted a selfie as your profile picture on this very site, Zico 😁

5 hours ago, only1swanny said:

Yes, they will..

Schools are now moving towards AI created images of students and teachers, ensuring that staff are not on the website, and any students featured are past pupils.

Social media posts are slowing down for some schools and featuring fewer images.

Teens are getting pictures from websites and using AI to put teachers (and sometimes other kids) in awkward situations. It's a crying shame as some kids and parents love having their kids on the schools media, showing excellence or achievements, but they can't run the risk. I saw a video of a former colleague sharing a bed with another male teacher, marriage and career wrecking stuff that. I've had a picture of me taken at the match and passed around and that's bad enough!

The amount of paperwork involved in taking pictures in a school is crazy. I'm one of the school photographers, when I take my camera in, there has to be a designated memory stick which is either in my camera, or locked away, and deleted as soon as they're on the school system. I took a brilliant picture of one kid at the school show, but I can't pass it on to the child's parents as it's got another kid in the background.

That’s mad

Aye it’s such a shame.

Not sure some dude is gonna knock one out over my lad posing with a football medal tbf like.

18 minutes ago, gonzo said:

Not sure some dude is gonna knock one out over my lad posing with a football medal tbf like.

That's what most parents think about first,

however the reality is there will be some cruel kids who use AI to create daft images, what starts as a joke can turn sour really quickly.

Then the really vile ones that use them to create fake videos.

3 minutes ago, only1swanny said:

That's what most parents think about first,

however the reality is there will be some cruel kids who use AI to create daft images, what starts as a joke can turn sour really quickly.

Then the really vile ones that use them to create fake videos.

Ah yeah never thought of that angle.

5 hours ago, gonzo said:

Not sure some dude is gonna knock one out over my lad posing with a football medal tbf like.

send me the pic and I'll decide

6 hours ago, gonzo said:

Not sure some dude is gonna knock one out over my lad posing with a football medal tbf like.

That's how I feel about it. Chances of something actually happening are insentismal - but we hype the threat of it up.

Paedogeddon style 😁

On the flip side of bad news stories. This is an amazing breakthrough

https://www.bbc.co.uk/news/articles/c5y3j3ngevmo

10 minutes ago, kent_white said:

That's how I feel about it. Chances of something actually happening are insentismal - but we hype the threat of it up.

Paedogeddon style 😁

"Instentismal" isn't a word.

32 minutes ago, kent_white said:

On the flip side of bad news stories. This is an amazing breakthrough

https://www.bbc.co.uk/news/articles/c5y3j3ngevmo

What’s the worst that could happen?

2 hours ago, green genie said:

What’s the worst that could happen?

Must admit, this news immediately had me contemplating the global over population problem which is behind much of the migration away from the area between the tropics towards (currently) more temperate climes.

Giving AI the wherewithal to design benign viruses then raises the spectre of AI going rogue (again?) and creating and finding a way to release malignant ones when it identifies the imbalance in the equation of resources versus population and decides to act. Potentially the end of the world as we know it?

I think I may just have conceived a new Hollywood blockbuster.