Can AI just fuck off please?

[noblegiraffe]

I've not been impressed with AI since I asked ChatGPT about noblegiraffe on MN and it told me that I'd voted for Nigel Farage and had right wing views.

But recently, it seems like every bloody thing has got AI sat on it. I just looked at my emails and the preview pane instead of showing the first line of the email now has a little précis of the email. I went to twitter to see what is being said about the Trump fiasco and every other tweet now has AI asking questions about the tweet at the bottom of it like a surprise English comprehension test (see below). I open a PDF and it asks me if I want a summary of the document. I went to send a text and it suggested I reply 'Sound's great!'.

NO. You are the modern equivalent of Clippy and no one liked Clippy either. Just Fuck Off.

It depends on the job. I used to be a freelance copywriter, having left teaching because I developed an autoimmune condition and am immunocompromised. My work allowed me to earn a good living flexibly around my illness, and I worked hard to upskill to change careers. Job opportunities quite literally disappeared overnight when Chat GPT arrived. It was absolutely horrendous and I’ve struggled to work out what on earth to do as I am ill so often that getting a ‘normal’ job is no simple thing.

I’m really one of the lucky ones because I still have some loyal clients, although making a living is a struggle now. Fortunately, I write in some niches that require extensive research from multiple sources, requiring me to make judgement calls to reconcile conflicts between different documents, which AI isn’t great at. It’s also deathly dull to write, so I never had loads of competition for the work pre-AI. Most of the writers I know very quickly had no work at all.

You can see the results of all this on many websites that previously put out well-written, informative content. Back in the day, it used to be low-quality websites putting out spun content (running plagiarised text through software to change enough words to similes to make the content ‘unique’). At least this was instantly obvious because it was almost unintelligible. Now, it’s loads of bland, formulaic stuff that seems ok at first glance but is actually written by a machine with no insight or understanding of appropriate tone, usually accompanied by weird AI images full of distortion.

If you use Google and type -AI when you search it doesn’t give the AI summary.

My boss got ChatGPT - the pay version - at work to be used by a couple of people at work who struggle with writing. We started playing about with it and get it to rewrite stuff and I was actually quite impressed. Used it again last week and it was absolutely shit. As someone pointed out- it learns by what people put in it. So 9 months of idiots using it has just made it stupider. I asked it to check grammar and spelling as a test and it missed loads of mistakes.

Once you know what you are looking for its really obvious. The use liberal use of the dash - is one obvious way. I think thats going to be the death of it and companies will actually start to prize well written content.

You can also use other search engines like DUckDuckGo or Ecosia. I like the latter as it runs a tree-planting programme.

Like I said on another thread, not using AI is fast becoming a selling point in my trade, which involves writing for a living.

The problem as ever is that we have embraced a tiny number of US apps and as they embrace AI we are pushed onto it. I have zero faith that it won't be used to influence social change and politics. The only way around it is for people to get permanently offline (not going to happen) or alternative platforms developed in UK or Europe are rolled out without AI. I know so little about tech but am thinking I might move away from Google to an alternative that doesn't use AI. My phone defaults to Google products so I don't even know how to not use it but I'm prepared to try. If for environmental reasons only we should be looking at this.

I can see the benefits of AI for improving efficiencies in back-end already automated processes (data processing etc) but as someone whose entire career has been, and still is, built on creating effective, honest, original, legally sound and accurate communication, Chat GPT and the like can bugger right off.
We were recruiting recently for a junior comms role, and had to run 150+ applications through an AI checker, as it became clear that the majority of candidates had used Chat GPT etc to create their responses.
I liken Chat GPT to the person in the pub who goes round sweeping up all the left-over dregs from other people's pints - crap in, crap out (maybe this was an 80s/90s thing though 😂)

@MargaritaPracticallyCan

'We were recruiting recently for a junior comms role, and had to run 150+ applications through an AI checker, as it became clear that the majority of candidates had used Chat GPT etc to create their responses.'

The irony...

I definitely do not excel in marathon based environments.

@Summerartwitch I know! The penny dropped when I saw the exact same sentence repeated on three different applications, with absolutely no attempt to edit it, blatant copy and paste - did they think we wouldn't check?!
Thankfully not all younger people are doing this - DS has just graduated, his diss was on the rise of big language models, how it impacts people entering the workplace and how damaging and limiting it is, for all the reasons everyone's pointed out in this thread.

If there are facts about your company it has no access to, it doesn't have the information it needs to make a sound decision for example.