Post

1. Recently I’ve received lots of emails from people I don’t know personally, asking for my feedback on their books or papers. They explain the connection between my work and theirs and ask me to take a look. The catch, of course, is that the emails are transparently, obviously, written by AI.

2. I guess this is the new form of academic engagement farming. Spin up some agents to search for relevant targets and then write each of them a letter, custom tailored to what the target has published, about whatever you’re trying to promote.

3. I find this a profoundly cynical thing to do. The academic community relies on norms that enable productive interaction among community members, even if they have not met previously. It has always been the case that I can assume that a fellow academic contacts me is doing so in good faith.

4. These emails violate that compact — and the consequences that they force me to stop doing something that I’ve always done for my colleagues. If someone I’ve never met writes me a long letter, asks me to look at their work, and it’s actually relevant, I do. But not anymore.

6. When I was 24 years old, I wrote a long letter to my academic hero, John Maynard Smith, asking him for comments on a paper I'd written about one of his models. He had never heard of me but my missive launched a (pen-and-paper) correspondence that ran hundreds of pages and lasted until his death.

7. It's yet another thing that AI has stolen away from the current generation of scientists. Yet again, AI proves to be an anti-egalitarian force in that these abuses force us to abandon the Mertonian norm of universalism for increased reliance on credentials, reputation, and personal connections.

I guess if someone were to sit down and write a letter to you with pen and paper, this would sufficiently stick out to be worthy of note. In an environment of frictionless communication, making the deliberate choice of investing personal time into something that is not frictionless is relevant.

A lot of the problems generated by AI (and email) can be solved by going back to using old technology. Writing a letter to academics is still possible, and if enough letters were sent we would be getting back our pigeonholes.