To worry that AI will alter the world beyond all recognition, and not for the better?

[Appalonia]

I know it's only in its infancy but it's already created music, screenplays, works of art, photos, essays, marketing, and has the ability to create so much else which will basically erase the necessity for so many jobs, including creative ones. What will pp DO when most of us are essentially irrelevant? And what will be the purpose of life when there's no point learning a skill when AI can do it so much better, and cheaper? I was watching an interview with Elon Musk who was saying that universal basic income is the only way forward. Which is great, but, if human beings are going to become essentially redundant ( apart from purely physical jobs such as cooking, building, hairdressing, beauticians, farming etc), what will be the point of us?

Just saw this on Twitter, short videos created by a simple prompt. They don't all look 100% real yet, but eventually they will. Sure, these are just fun, but the potential for mayhem is huge...

https://twitter.com/AJEnglish/status/1652417219404636161?s=19

It is going to change things - perhaps as profoundly as the internet has. But every time people have feared so far that the jobs will all be replaced by machines (or indeed by outsourcing) this has come to pass on a far, far smaller scale than was thought. There were lots of predictions that by now we'd all be being served by robot wait staff and being cared for by robot nurses and that has very much not happened, and not really because the technology isn't there yet. People are willing to pay a premium for stuff made by people. Buying art isn't just about ending up with a thing that you think looks nice, because if it were then artists would have been made redundant once we could make high-quality copies. There's a novelty value but beyond that people don't want computer art. Jobs will change and some jobs will go, just as they have with previous new technologies, but there will be lots of new ones.

I don't know why Elon Musk and many others signed a petition to try and curb AI , but it worried me a bit that they felt the need to.
Someone suggested on another thread it was because they want to keep the money rolling in to them , but surely inventing new AI or making the existing ones more advanced would be beneficial to people like Musk? It didn't make sense to me.
I hate the chat messages you get online from utility companies etc and often you end up having to speak to a human anyway if it's a more complex problem. It sometimes sends you in a circle or doesn't understand certain words or how we pronounce words etc ( if it's a spoken AI recording for bank passwords or something like that)
Most people would rather speak to a human from the start

The Internet has already altered the world beyond all recognition.

i’m a lecturer and it’s becoming increasingly difficult to set assignments that can’t be done at least partially to quite a high standard on ChatGPT. And there is no way of really detecting it at the moment. Degrees and other qualifications will become meaningless if we’re not careful.

We all know what happened when Skynet became self aware …

The big problem I see is AI won't be regulated. The Internet has been around for a while now and we still haven't found a way to regulate that either. And it's undoubtedly changed how people behave in many ways. I remember work telling us not to worry about AI because it will be regulated. But I really can't see how it could be. The potential for even more misinformation is absolutely huge.

When Tom Berners-Lee gave the Reith lecture, he said the internet could be regulated but no is interested in doing it.

AI will be the same.

That should say Tim
and no one

My snapchat has a new friend added, an AI one, whom cannot be deleted. Very creepy 😱

Can you report the account to Snapchat?

Crispr sounds utterly terrifying.

Because this is definitely not how we get terminators…

Yanbu
There's an awful lot of science fiction that could easily become non-fiction. Eg Skynet in the Terminator films

Crispr is terrifying… but exciting too.

I read an interview in nature with the Chinese dr who has just finished his two year sentence for illegally (but with the parents full consent) attempting to edit unborn babies genes to be immune to AIDS

horrifying as we find it now I bet in a decade or so we will find the idea that people wouldn’t crispr their unborn fetus to be immune to disease as shocking and repellent as most of us find the idea of not vaccinating our children now.

What so you want them to suffer needlessly of diseases that are for most something of history books?

Something or other is always altering the world beyind all recognition. The traditional "big 3" were gunpowder, printing and the protestant religion. AI kaes it 4.

No point in worrying.

The real question (which was part of my degree a lifetime ago) is what will happen when AI exceeds and surpasses human intelligence ?

AI designing AI is where we are going.

Oddly very few people worried about lathes making lathes (and other machines).

If nothing else all this talk of AI might just help us define intelligence a bit better. As the moment it's very much in the "I can't define it, but I know what it looks like" realm.

makes it

Maths and accountancy degrees didn't become obsolete when someone invented the spreadsheet. Universities need to adapt to take account of new technology not fight against it. We need to be valuing more advanced individual thinking rather than rather than fighting against the inevitable technology. The days of credit for memorising regurgitating information are over (and should have been a long time ago).

There is a whole new career of "AI wrangle" emerging. Because you can't just trust AI (remember, we don't really trust humans).

In my case it's wading through volumes of code and removing the crap. Which is usually some section that uses a fantastic - and totally fictional - system call or library function.

A colleagues said it was like dealing with "that mate" who knew that "vaccines give you cancer" and whose evidence was "google it".

I can spot shit code a mile off. I presume at the same time there are doctors who can spot shit diagnoses a mile off, and engineers who know that bridge won't stand. Because if not, we are in for an interesting future.

Apparently you can ignore it but not actually delete it 🙄 you can make it invisible but it will still be running in the background

@Easterbunnywashere I quite agree, which is why we don’t do exams which “regurgitate” information (and which, ironically, are the most immune to ai interference if done in person). We use methods that replicate the work environment for our discipline - this is easier to work round at the moment in terms of AI with quantitative / data skills, but less when you are wanting to assess more substantive issues using text-based methods. If you have any specific ideas though please do share.

The Snapchat bot is literally meant to be there, isn't it? They want you to treat this as one of your friends. It will track you for everything. Get rid of Snapchat.

Luddites have always been around to complain against progress. Automation has been happening since the dawn of time, why the hysterics over jobs being lost to automation.

On the UBI point, sadly we’re half way there already. In this country alone, there are 6 million people on universal credit. That is a sizeable chunk of the working population who depend on the ever dwindling number of net contributors. Unfortunately there are three classes being formed - a ruling class that spouts this nonsense about the need for UBI, those that pay for the privilege of the ruling classes to spout the nonsense and then the ever growing vegetative class at the receiving end of charity.

It's already changed the world, but...that's the initial bump. It's not going to be as scary as people are making out, it's just a tool. There's also the fact that it requires massive amounts of human-generated training data in order to function.

Thing is, much has been made of GPT-4's ability to write code, but most people don't really understand it - it might be able to write code, but it can't write great code. In fact, in order to get it to write good, readable, performant code you have to put more effort in than it would take to write it yourself...and, critically, in order to get it to do that, you already have to know how to do it yourself.

The result of this will be similar to the old days of business functions being written in Excel and Access with VBA - lots of people churning out crap code that eventually needs to be sorted out by someone who knows what they're doing if it turns out that it's actually useful.

It's also ridiculously confident in its wrong answers, and the more complex the concepts you give it...the more often you get a wrong answer.

There's nothing to be worried about, as long as the people using it know its limitations. The worst case scenario, as things stand, is lots of crap being put out by people who don't understand the questions they're asking it.