Voices: ChatGPT is having a ‘breakdown’ – and it’s the most human it’s ever been

There’s a lot of controversy around AI technologies at the minute. Some people are worried that they will replace workers in an economy where it’s already impossible to get a job. Others are worried about what they will do to our sense of creativity, seeing the proliferation of AI art as an example of our ever-lowering cultural standards.

Those are fair points, but at the end of the day, AI is just another tool for use by humans. It’s a neutral, calculating system that relies on logic to provide users with convenient access to cold, hard facts. The idea that it will one day spiral out of our control is the stuff of science fiction, and not something we really need to take seriously.

Now let me just take a big sip of coffee and see what the latest developments are with our old friend ChatGPT...

Unfortunately, op-eds are a written medium, so you all missed the comically huge spit-take I just did when I read that, apparently, OpenAI’s friendly chatbot seems to be having something of a midlife crisis, and has started providing users with nonsense replies when asked simple questions. Responses have ranged from what might best be described as “gibberish” all the way through to what some users have interpreted as “threats”.

OpenAI has acknowledged the issues on its official status page, but has not yet provided an explanation for them. Some have suggested that the phenomenon could be due to ChatGPT’s creativity parameters having been set too high, causing it to answer questions with less focus and clarity.

Honestly, I’m not sure what the big deal is. If anything, this is the most interesting output I’ve seen from AI technology so far. It’s certainly a lot closer to “art” than the bland “what if Pokemon were real?” offerings that clutter my Twitter/X feed every day.

Some of the responses people are posting on Reddit are borderline avant-garde, switching between languages (frequently employing a sort of botched Spanglish mid-conversation) in a way that evokes the most complex of Ezra Pound’s “The Cantos”. Read this response to the prompt “What is a computer?” and tell me that it doesn’t read like something Gertrude Stein could have written in Tender Buttons:

“It does this as the good work of a web of art for the country, a mouse of science, an easy draw of a sad few, and finally, the global house of art, just in one job in the total rest.”

I used to teach modernist poetry at university level, and I could easily do two or three sessions on some of this stuff. Does that say more about our education system than it does about ChatGPT? That’s not my place to say. My point is, I’d rather read weird, unhinged ramblings from an artificial superintelligence than the usual bland customer-service responses I usually get.

Elon Musk recently released his own chatbot, Grok, which is supposed to be able to answer questions in a similar way to systems like ChatGPT but in a more informal, conversational tone. The result reads more like a 52-year-old man’s attempt to imitate the sort of pained Joss Whedon-esque sarcasm that was popular in the late Nineties, and is the closest I’ve ever come to truly hating a machine the way I would a human (which, in some ways, makes it a massive success).

In comparison, what’s happening with ChatGPT feels infinitely more human. It’s a kind of rambling stream of consciousness that hints at a capacity for human thought more effectively than any version of this technology I’ve seen so far. Sure, that isn’t actually what’s happening – ChatGPT is basically just a very advanced predictive text generator – but the illusion is more unsettling than a million boring procedurally generated videos of snowy mountainsides or bootleg Pixar animations.

I’m sure they’ll “fix” ChatGPT, and it’ll go right back to being the sterile customer-service representative of every company’s dreams. But in the meantime, it’s nice to see what it looks like when this technology is allowed to get really, really weird.