To be so sick of AI!

[Missingpate]

It’s everywhere and it’s awful! On a personal level, taking over my FB feed with nonsense, fake images that are just annoying, irritating platitudes… the first thing you get on an internet search is a load of Ai content scraped from all over so you have no idea what you are reading and can’t begin to credit a source if researching etc. it’s basically theft, as I see many are now petitioning to point out.
And it’s terrifying for what it is likely to do to the jobs market and all sorts of other Black Mirror type things my brain can’t fully imagine but I’ve no doubt are coming. Yesterday my boss had an email from a company with an AI plan to do what I do, explaining all the things they could offer. We all had a good chuckle about it because it totally lost all the nuance of the actual role involved but for many this will become reality. It all feels nightmarish quite frankly. I’m sure there are some good things but struggling to see many. Medical ones come to mind, I suppose. But wanted to rant, so ranting 😂

But she had to tell the AI what to say so I don't see the problem here. She's probably thrilled that something that was hard for her - to explain how she feels living in her new house - came out better than anything she could have done herself.

Just like ghost writers write autobiographies. Or those articles/emails/newsletters penned by the CEO are not actually written by the CEO.

I helped a friend write a letter recently to her child's school. She's not English, is really frustrated by things that are happening there and explained it all to me, but when she tried to put it in an email, it was really hard for her. I'd like to think the effort I put in represented what she wanted to say but more clearly and eloquently than she would have been able to say it herself.

I would definitely support that if it could be guaranteed in some way! AI plagiarises and its generative writing cannot be trusted to be accurate.

Sorry about your career. I've written a few science books too, but doubt I'll write another as AI can do it faster and cheaper.

This also drives me mad! I don't need you to summarise it - I need you to move so I can read it!

So many of these posts are just nonsensical to me.

We've been getting shitty articles and commentary on facebook practically since it was invented. I feel like the first 5 years I was on Facebook I spent most my time commenting on dire warnings and the like with endless links to Snopes pointing out how untrue they were (no, babies are NOT being grabbed by mutated killer foxes. No, if someone flashes their lights at you it does not mean you're about to be murdered in a crazy gang initiation. No, eating raw onion will NOT kill you etc).

Social media, technology, globalisation etc have definitely had the effect that lots of people with few or poor critical thinking skills can now boost the dumb things they already thought, and interact with loads of other people who are just as dumb. So that's a problem, yes. But that's not an AI problem specifically. In fact, in theory, AI could solve for some of this if the smarter people who are training the AI are doing so on legitimate sources.

There's a vast difference between humans teaching and supporting each other with a degree of community and the understanding that there exists nuance and fallibility in such exchanges. These are organic processes that teach us about ourselves and other people and how we can work together.

The danger with AI is the encouragement to believe it is infallible and superior to human input. In addition it makes the idea of accountability when things go wrong far more complex. On a very basic level the Post Office scandal illustrates this in spades. More faith was allegedly put in technology than end users raising the alarm. Actually it was a cynical exercise in avoiding accountability that finally, thanks to human diligence and persistence, was brought into the open.

For years corporations and institutions have been using "computer says no" or "computer made me do it" as cover for inefficiency and the operation of inbuilt bias towards certain sections of society.

Anyone who can't keep up with the ever increasing pace of technological progression or has found themselves out of a job or blocked from some process by a skewed algorithm is just chided for being a Luddite and expected to "get with the programme" or accept societal disadvantage and exclusion in extreme cases.

And don't get me started on some of the ideological underpinning behind the tech bro mafia. Such as encouraging certain people to breed more prolifically than others. There's a name for that.

@MrSeptember Yes but I think you miss the point. Yes we’ve always had a lot of stupidity on the Internet. This is about how much content is now generated by AI. It’s everywhere - making most of the large Internet sites unusable (don’t get me started on LinkedIn.. )

This video is quite timely..

They call it the “ensh*ttification” of the Internet

The Gartner hype cycle very much applies here in my industry - loads of excitement but very little good output which is leading to disillusionment!!

I was messing about on my phone the other day and an ad came up for AI boyfriends. I don't know that much about AI so I clicked it out of curiosity. Ended up "chatting" to this AI. I said I had 2 cats and it told me I should get rid of one as it would make my life easier! The romance was over at that point 😂

It clearly still has a long way to go in that regard at least. On the one hand, I can see that something like that, has the potential to help people who are lonely and don't have anyone to talk to. On the other, some people will end up avoiding human interactions and relying solely on AI.

It's like anything new, there's going to be good points and bad.

@MrSeptember

I helped a friend write a letter recently to her child's school. She's not English, is really frustrated by things that are happening there and explained it all to me, but when she tried to put it in an email, it was really hard for her. I'd like to think the effort I put in represented what she wanted to say but more clearly and eloquently than she would have been able to say it herself.

This is a perfect example of the sort of act of human connection and kindness that will be lost when people turn to AI instead.

Yes, having AI write that letter would have been quicker and easier. But I value many, many things above speed.

It is all part of the march towards a colder world with ever-diminishing opportunities for human engagement. Some people are excited by the efficiencies such a world offers. I personally think we are giving up far too much in exchange for those efficiencies.

I haven't missed the point. Stupid content, whether generated by AI or not, is stupid content and has existed for a very long time. Stupid people who accept all this content are not new. If anything, AI has the potential to limit some of this stupidity. But of course, it assumes people are smart enough to use good AI....

AI is here to stay, the technology won't go away. And it's in its exponential phase so it does seem like it's everywhere at the minute.
I use it daily for my personal life and my job so I love it, but I don't like how posts on Facebook or Mumsnet aren't labelled as AI. It's very misleading.

Interestingly my teens are very good at identifying AI and critically evaluating it. A skill older generations need to learn!!

Sorry about your career. I've written a few science books too, but doubt I'll write another as AI can do it faster and cheaper.

My DH has written a couple of specialist science books.
I can't foresee an AI any time soon being able to assemble and sensibly arrange the relevant, correct, properly referenced material any time soon. There may be ways they can assist in the process but I wouldn't trust one any time soon.

AI has totally ruined Google Images - just a few years ago you could find beautiful digital art that was made by a human. Now the same search results turn up mass-produced, low quality AI nonsense.

I used to share things I found on Facebook but information and pictures can no longer be trusted at all so I don't bother. It's also becoming increasingly apparent that Facebook has more bot accounts than real people and a huge proportion of political posts are created by AI and copy-pasted to multiple different bot accounts.

And that's without going into how insulting AI feels as an artist who's worked very hard to reach her current skill level.

The sooner AI wipes out all jobs and we're put on a universal basic income the better. Lots of people hate their job. For many, it grinds the life and joy out of them. I went into a cafe the other day, and the woman (who has worked there for years) looked so tired and bored, so utterly sick of her job, that I almost felt guilty for ordering. It's the same for the staff in my local Sainsbury's. They just don't want to be there. I remember going in there on a hot, sunny day, and the guy in front of me at the fish counter said to the woman "it must be lovely to be in here in the cool." She gave him a look like she wanted to kill him and said "if one more customer says that to me I think I'm going to scream. Do you really think I want to be trapped in here on a day like this?" That's how a lot of people see their job – as a prison.

Yes, I would expect most writers in this sort of scenario to be using AI. To check for logic errors, to clarify wording, to offer suggestions etc. And different AIs are better at different things. I've been using Chat GPT a lot but it's useful fo rresearch, terrible for any sort of content. I've heard Claude is good for content help.

Having said all that, the value of AI to science and medicine is clearly enormous. We are going to see a pace of progress there the like of which has never been seen before.

Yes - I'm nearing retirement now from a long career in scientific software development. There's a lot of exciting stuff going on, lots of new hires too. And a Nobel Prize

www.nobelprize.org/prizes/chemistry/2024/press-release/

Developing and using this stuff takes a heck of a lot of human intelligence, and it's built on huge amounts of experimental scientific data. It feeds back into physical research - so, use an ai to better understand something, maybe design something new - but then you have to actually make and test it in the real world.

What I'm rapidly discovering is that as a discipline, machine learning (A.I., language generation tool, whatever you want to call it) needs the humanities. Machine opacity is one serious concern, but this discipline simply cannot operate without ethics - philosophical ethics, that is; not simply passing specific universities' ethics committee as every research project completed under their banner must do. Thus, everything in its programming needs screening for four factors: responsibility and accountability; reliability and safety; fairness and inclusion, and privacy and data governance.

And here's one on intelligent machines making 'ethical' decisions; a scenario very much based around Philippa Foote's 'trolley' problem in determining utilitarianism (the greater good for the greater number):

‘The traditional example is that of an imminent car accident involving an autonomous car. The onboard software has to decide whether to veer right, and face 100 percent chances of killing a young girl, or veer left and, for instance, having 50 percent chances of killing an elderly couple. [….] what should the autonomous car do? Its actions will be driven by the algorithm and the data, but this still poses a question: whether to base the decision on gender, age, probabilities, number of fatalities, their contribution to society or on any other consideration (Pellegrino & Kelly, 'Intelligent Machines and the Growing Importance of Ethics).

If you have machines deciding who gets to live and who dies in this way, you have some very frightening similarities with eugenics. Who decides what are desirable characteristics; who gives to live and die in a society; who contributes and who doesn't? You there have something more akin with the principles of a Nazi genocide. And whilst I don't necessarily accept that a hideous Skynet or Machine Wars or the prospect of being taken over by sentient machines is an imminent (or even distant) prospect a la Terminator or Matrix franchises, these are all deeply worrying developments.

I get that these are asides from data-scraping and a few irritating pop-up bots. But this thing needs very strict regulation, IMO, and by some democratically accountable process. Imagine the 'wrong', small, privileged group of people having too much control over the programming of this tool. Here we'd have a media monopoly with far too much influence over what people (who are as a species impressionable) might think. And at the best of times, 'ethics' are a murky customer.

The worst thing is that bit by bit AI encourages us not to bother thinking.

"Don't think about the basic building blocks of your argument, don't research all the aspects of what might be involved, don't bother studying or actually learning, just learn how to use ChatGPT."

First we all develop a sort of attention deficit disorder due to quick swipe social media, then we outsource our thinking to a high tech plagiarism machine that has just nicked our work off us and monetized it back at us. I can't believe that more people can't see where this is taking us.

And the idea of a "universal basic income"? Yeah. As if.

Point me to an example from actual human history where the economic winners subsidised the economic losers.

"AI" is lots of different things, but broadly the overapplication and belief in it as a solution is misguided and ultimately a dead end.

Have you seen Pinterest recently? AI generated images of hairstyles, interiors and models - it's really hard to find something real and therefore achievable. It renders the entire site a bit pointless. And I hate the AI generated summary on Google now, even if it does indicate the sources it's using with the blue text.

I am hoping that the mediocre results for enormous expense (Google's nuclear powered AI search anyone?) will mean it's general application eventually peters out, except for technical and scientific use as @DaemonMoon points out.

It's pretty awful and reductive in terms of human interaction, recruitment and knowledge exchange, but people do really need to get their heads round what it does and what it can't do and how it does it. It is, as they say, neither artifical nor intelligent - it's computational patterns at scale and often totally dislocated from source and context. And this is before we get into all the myriad ethical issues - what looks like no-work solutions for you is actually reliant on hidden and highly dubious practices labour elsewhere.

The sooner AI wipes out all jobs and we're put on a universal basic income the better

I wish I shared your confidence that this will be the outcome!

My fear is that we will instead see similar effects to what happened with increased use of automation and overseas outsourcing in manufacturing. Widespread unemployment, more and more people competing for fewer and fewer jobs, people left bereft of meaning and in appalling economic deprivation after their skills become obsolete almost overnight.

This time it will affect the middle classes/office class, whereas previously it was the skilled working classes.

Perhaps eventually we will move to universal basic income and freedom from the prison of work! But I’m afraid it may be very painful getting there.

The basic income will be enough to keep us as mindless consumers but can’t see the tech billionaires sharing my more than they need to!

Have you seen Pinterest recently? AI generated images of hairstyles, interiors and models - it's really hard to find something real and therefore achievable. It renders the entire site a bit pointless. And I hate the AI generated summary on Google now, even if it does indicate the sources it's using with the blue text.

But this is what I don't understand. That's just not my experience. I was just on Pinterest last week looking up easy to make halloween snacks for a party we're throwing. Lots of great, easily achievable ideas - DD and I have compiled a list ahead of doing the shopping next week. So what am I doing that is generating content that is either not AI or AI-but appropriate vs what other people are doing?

Like any tool or technology, the way it is implemented and used is essential to the process.

I write more popular science books (basic biology/science with lots of pictures), and the publisher typically has a scientist write them and then various consultants check them. There's not much to stop them having AI write them and then consultants check them, especially for big businesses focussed mainly on £.

I've recently retired and one of the last things I had to do was g through the Google listingfor over 400 branches and individually switch off AI updates as it was putting in absolutely bonkers opening hours. Eg teeny branch in teeny town open 24/7 or a branch in a shopping centre that's not accessible overnight open 8pm to 6am.