Stephen Downes

This is a longish but useful article on the state of play for the use of AI in writing academic articles, including examples from journals and surveys of journal policies. In a nutshell: it's mostly OK for some things, but needs to be declared, and a human must always be responsible for the work. The suggestion is that a human needs to do the intellectual work, so while an AI might be useful as an assistant, it must be a human that is doing things like forming hypotheses or interpreting results. I think these boundaries too will shift as AIs take on tasks too complex for humans, in which case the proper form will be something like "the xyz AI interprets these results as abc".