functional transcendence

Baudrillard, automation, and ChatGPT writing

I just want to say a quick thanks to paid subscribers—new and long-term—for your support, which increasingly allows me to spend longer writing these essays, digger deeper into topics and hopefully providing more thoughtful analysis and better researched pieces. Being able to revisit topics like the one below is not only creatively engaging but is actually changing how I think—so thank you! And do let me know if there’s something you’d like me to write about here.

---

In the last couple of months, many of the writers I know—the ones with contact forms on their websites, at least—have been receiving the same sorts of emails, up to several times per day. These emails are extremely flattering, weirdly formal yet falsely friendly, and almost identically structured. They run over the details of one of your books, cribbed from marketing materials or public reviews or a combination of these, and throw in a creepy personal detail as well—before going on to try and sell you a marketing package or hook you into an ‘opportunity’ to go on some (really existing) mega podcast or TV show to talk about your book, for reasons I can’t really ascertain but am obviously not going to chase. They read like the sort of faux-inspirational, excessively-paragraphed slop you’d find on LinkedIn (and, increasingly, Instagram) from people who style themselves as ‘business coaches’. Here’s one such email, sent to me last month:

I’m not sure why naming my cats Ernesto and Fidel makes me fearless (in fact they were going to be called Worker and Parasite; we chickened out), but that’s the least of the issues with this. Perhaps if you had just fallen off the turnip truck you might be fooled by such a communication, but for most of us this is so obviously ChatGPT-generated that it’s grating to an extreme degree. If you’re going to try and scam me, at least have the decency to write the template email yourself.

Of course, scam emails are a thing you have to deal with when you have a minor public presence and an accessible contact form. The frequency of them has massively increased, but that’s not the end of the world. The issue is that we also receive genuine emails through those contact forms, emails from complete strangers asking us to do things or offering opportunities. And those messages are also increasingly written via ChatGPT, to the exact same beats, sentence structures and tone as the email shown above.

The result of this is that almost every cold email is being considered as a scam email and dismissed out of hand. How does one tell the genuine apart from the fake, when the two are borderline indistinguishable? You just don’t bother. A human-written email will jump out from amongst this, if email filters don’t funnel it automatically into spam or marketing because it’s come via the same route as all this other trash. But the sheer tide of spam means that a lot of people will close their direct methods of communication and force all requests to go through their agent or publisher, who have even less time to try and detect the real emails from the rubbish. Even if these real emails do get through to the intended person, it takes such an effort to try to parse meaning (and intention) from the generated text that the recipient becomes annoyed at what a massive waste of time it has been, leaving them not exactly amenable to working with the sender; if every email is going to be so inhumanly written, why would you engage?

If you are a person using ChatGPT to write these requests, and ‘saving yourself time’ by doing so, the grand-scheme effect is that it is making your emails less likely to actually reach the person you are trying to get to, and more likely to be ignored when they do. Outsourcing communication to technology, it turns out, makes all communication sound the same, and those on the receiving end have to put up further filters to avoid being overwhelmed by it. In all the endless discussions we’ve had about whether AI would ever be able to convince us it was a person, hardly any considered that actually, people would just willingly make themselves sound like AI instead. The very thing that promised to ease human interaction is making it more difficult, in the end.