Off Topic A.I. will be the end of all of us...

Catchy tune

 

I've started singing the chorus to myself already to be fair

 

I am leading a large digital transformatiom programme currently (posh title for better use of IT). AI is a biggish part of it, but not a really big part of it to be fair. We have settled, quite confidently, on the view AI cannot replace human intelligence. We do believe it can augment it though. We talk about using AI to augment humans, never reduce of replace. This is supported a lot by emerging research about where the boundaries of AI sit.

Some examples. We use a lot of AI in automation of workload. So we take task that would be done by humans, sat in an office, or a workshop, that are really low value tasks. We at the same time have people then deciding what high value tasks this now free time can be used to do. In my sector that is nearly always proving to be talking to people / customers. It is amazing when you go and talk to your customers, who in my case are often younh people, they will tell you what would make their experience better is seeing and talking to someone else about a problem or a thing, who has empathy with it. Generative chat, like ChatGPT, falls short in a lot of areas. It is very good for some things, but in a real world chat it is miles off what a human can do when chatting. It has no sense of emotion or body language. It can only deal in out of date facts currently. There are definite uses, but again only ever to augment us, not to replace us.

Personally I think AI is nothing to worry about. It will have a place, but the human brain will always be several steps ahead. We will adapt to these capabilities, use them, and find ways for humans to add greater value. With any luck it will allow us more time to create things of real beauty again. I was at Stratford and watched Julius Caesar on Saturday and no AI can produce the sort of joy or delight that can.

 

I can see it becoming an absolutely massive problem. The advances in the last few months have been absolutely incredible. Within a decade I can easily see it being able to do pretty much any office job in the world better than any human, in a fraction of the time. So, why would companies continue to hire humans?

Suddenly, half of the world's workforce is unemployed and all their years of training / expertise is now worthless. This could easily lead to rioting on an unprecedented scale, not to mention the economic impacts of massive unemployment. I don't think it's an exaggeration to say AI could easily result in the collapse of the global economy as we know it.

And this isn't even considering the threat of a "Terminator" / "Matrix" style outcome, which is now a 100% genuine possibility.

Honestly, if you think about it too long it becomes absolutely terrifying.

 

AI can't make the intuitive leap, or connect apparently disconnected elements. That might change when we get to quantum computing where it can brute force an approximation of it.

 

2 mates of Mine are working on a podcast, there's something similar using voice. you put whatever you want it to say into the system, then the next thing you know, Samuel L Jackson is reading aloud your paragraph. It is used for good, but it could end up getting used for evil. People in power ie President will probably need more code words when giving orders as it'll get to a point where a you'll not know what's real or not. All I'll say is with A.I, Terminator 1984 tried to warn us.

 

Totally agree with all of this. I've always said, if a machine replaces a person, be the person that fixes the machine.

 

It's scary. Listen to this:



There is absolutely no way to tell that is not Jay Z on the track, u less you are told. So it will take over creative roles as well. There has been a huge uproar in the digital art world as AI is essentially making artists obsolete. Initially you could tell what was A.I, as it couldn't get hands or teeth right, etc. but it has learned how to draw them incredibly quickly and is only getting better with each passing day.

 

Unreal that. A.I could finally bring the world to its knees, by helping people smash exams, give them qualifications in fields they know nothing about, getting them a top job in a company they have no know knowledge in, and once these chancers are in, the damage begins. Its not too dissimilar to when they handed out predicted and not actual high grades to students in 2021 like sweets. Going forward when they're sat there in 5 years time in an interview, "so I see you're were a straight A student?, what year did you graduate? ...."erm, 2021".... Oh.

 

They're already there, tbf. Individuals who've blagged all the way up, played the system for all it's worth, then end up in a position where there's nobody above them as a guide and the ****-ups begin.

 

Can the person who is being impersonated say that’s not me the same as you do now?

 

If you get the chance, read “The Future is Faster than you Think” by Peter H Diammandis and Steven Kotler . There’s a lot of stuff nearer than we know about , often just infrastructure which holds it back ( eg flying cars , no ownership , just order a ride from your phone- nearest available will arrive , computers powerful enough to control millions of these in the air without collision etc)
Order online by simply giving a thought order
Loads of stuff including medical diagnosis and resolution.
Fascinating, but possibly scary as well

 

On the air traffic control bit - it can't account for human stupidity. Turning right when they should turn left, mishearing a level instruction, doing something without warning. The system is reliant on parameters that can occasionally be exceeded by halfwits.

If it's all tied into the one system without any potential for human interference it'll work. Would many of us entrust our safety to a machine we can't take over from though.

 

It’ll have no human control at all - car turns up , you jump in , sit back and get taken to your destination - even more advanced than The Jetsons !

 

Just read an article about this.

The fear seems to be godlike Artifical General Intelligence, which can basically teach itself without human help. So humans become redundant?

So no more writing essays for students, a macine creating music and being the best chess player ever?

However i am very sure i do not understand the implications of this.

The Internet has resulted in high street shops and banks closing down, cash being redundant. Porn on tap and conspiracy nutters finding friends.

At 3 score years and 10, should I be worried about this?

 

If only AI could help us look back a couple of pages to stop the scourge of repeated threads on forums


https://www.not606.com/threads/a-i-will-be-the-end-of-all-of-us.406296/page-2#post-16626788

 

Honestly, I would say dont worry. Nowt to do with age either. There is a lot being written about the dangers of AI currently and chat gpt has amplified it. But I dont really see some of the doomsday scenarios occuring at all.

Things like ChatGPT are flawed as it stands. That is because it is only as good as the data it has access to. By extension if data is crap then chatgpt will be flawed and we all know how much rubbish and incorrect data is out there. The trendy term is that AI can hallucinate, because it thinks it has found the right data to use, but it hasnt. I was talking to an expert out in the US last week, who has spent years researching AI. He was telling me about the hallucination issue being a real puzzle.

Artificial general intelligence I doubt is close, if even possible. I tend to think we wont see anything like the sci-fi movie scenarios. Some researchers suggest AGI is decades away at least and even then it is hard to see enough progress to see us humans ousted. My opinion anyway.

 

Interesting, mate. Dont you think chat gpt is more a proof of concept? If they trained the model on say, all of the significant medical research that we have, how could a GP contend with that level of computing power? I've been in many an appointment when the GP has basically googled my symptoms in my presence. A medical A.I that could genuinely learn would make GPs redundant pretty quickly, wouldn't it? Same could be applied to law. A lot of the work they do is researching precedents.