About AI

In the New York Times, the neuroscientist Eric Hoel argues that the increased use of artificial intelligence is forcing any notion of intellectual or aesthetic quality into a death spiral, prompted as much as anything by human laziness. For example he refers to researchers at a conference on AI using AI to conduct peer reviews on AI-related papers, taking any human critical intervention out of the equation. Which is a problem, because one thing AI is very bad at detecting is bullshit, which is ultimately what peer review is for.

Of course, most of us don’t hang around at AI conferences, but Hoel suggests that the process is far more prevalent than that, eroding the fabric of culture itself, to the detriment even of people who reach for their weapons when they hear the word:

Isn’t it possible that human culture contains within it cognitive micronutrients — things like cohesive sentences, narrations and character continuity — that developing brains need? 

In other words, the processes by which people engage with all the gubbins of society is as significant as the content itself, and that’s what AI is stripping away. But it’s not as if the purveyors of AI are doing this deliberately, is it? They’re not consciously proposing policies that will make humanity that bit more stupid are they oh wait hang on...

PS: And even if you’re not that bothered about AI destroying the canon of Western literature, you might want to know what it’s doing to your fridge. 

PPS: And, following on from Musk’s tweet, I think this is supposed to be an April Fool’s gag but these days, who knows?

About the Beatles

I asked ChatGPT to identify the worst Beatles song and this is what happened:

It’s important to remember that opinions on music are highly subjective, and what one person considers the “worst” song by The Beatles, another might enjoy. The Beatles are widely regarded as one of the greatest and most influential bands in the history of music, and their catalog includes a vast array of songs spanning different genres and styles. Rather than singling out a specific song as the “worst,” it's more constructive to acknowledge that The Beatles had a few songs that were less popular or less critically acclaimed than their most iconic tracks. However, even these less celebrated songs have their own merits and are appreciated by some fans. Musical taste is personal, so what one person may dislike, another may find enjoyable. Ultimately, The Beatles’ legacy is defined by their many legendary and groundbreaking songs that have had a profound impact on the music industry, and their contributions to music far outweigh any individual song's perceived shortcomings.

There are all sorts of rumours that AI will take over from yer actual meat-and-mucus critics but this takes things in an even more worrying direction – it’s effectively decided that critical faculties are a bit impolite, actually, because you might be casting aspersions on somebody’s favourite. So it’s more “constructive” to acknowledge that some of the songs wowed the critics less than others did, provided you remember to assert the Fabs’ absolute centrality to the canon.

And it’s an easy question anyway. The answer‘s ‘What’s the New Mary Jane’, isn’t it?

PS: Elif Batuman asks ChatGPT a question about Proust and is told to, er, read Proust.

PPS: In American Songwriter, Jacob Uitti uses AI to imagine a Dylan/Cohen collaboration and, guess what, it’s dreadful. That said, much of the content here is as bad, even when it’s nominally written by humans. In a discussion of Paranoid Android we are blessed with this gem: “Despite the hurdle posed by censorship, the video managed to retain its audience’s captivation.” When I am king, bad writers, human or digital, will be first against the wall.

About AI

We should no longer be surprised that Artificial Intelligence is generating much of what we are encouraged to call “content”, whether it’s words or pictures (ceci n’est pas your mum). The tipping point comes when it’s not just the product, it’s the consumers who exist beyond meatspace. As Max Read reports in New York magazine:

Studies generally suggest that, year after year, less than 60 percent of web traffic is human; some years, according to some researchers, a healthy majority of it is bot. For a period of time in 2013, the Times reported this year, a full half of YouTube traffic was “bots masquerading as people,” a portion so high that employees feared an inflection point after which YouTube’s systems for detecting fraudulent traffic would begin to regard bot traffic as real and human traffic as fake. They called this hypothetical event “the Inversion.”

Which raises all sorts of questions: can there be a valid Turing test if neither party is human (but each assumes the other to be)? And does Baudrillard’s hyperreality become hyper fraudulent? (“Wasn’t it always?” chuckle the cynics.) And if we’re not brains in vats, could we just be phones in racks?