Science editor's "Slower, worse" AI piece used ChatGPT, essay says

The essay frames AI entryism as a long-term strategy in which pro-AI actors join or influence institutions while appearing to share their goals, or even while presenting themselves as firmly opposed to AI. It argues the weakest and likely easiest form is spreading rhetoric or cultish language that gets organizations to adopt harmful concepts through vague or damaging terminology.

The Science episode is presented as a case of that pattern: a figure with real power over academic careers writes a piece critical of AI, is found to have used ChatGPT in its references, and admits the use once confronted. The essay says some researchers with less power may now be less willing to voice objections to AI if they hope to be published in Science, and that this kind of influence radiates outward from senior people.

The author pairs this with a broader warning against judging by surface signals. Quoting Marx that all science would be superfluous if outward appearance and essence directly coincided, the essay argues that someone claiming to be opposed to AI is not thereby proved to be sincere — the same correlationist logic it says underlies racism and sexism. Its conclusion is that actions, not stated positions, should carry the weight.