This page records the journal’s policies on the use of artificial intelligence, both for authors and for reviewers. We aim to be consistent with the Committee on Publication Ethics (COPE), and more specifically with the ACM policies on AI in authorship and reviewing. The overarching principle is that AI tools for content creation may provide helpful assistance for scholarly processes, but cannot substitute for them.
We distinguish between AI used in generative contexts (such as large language models) that create content from scratch using only a prompt, and more limited writing tools (such as spelling correction, grammar checking, and language translation) that act on content previously created by a human author.
The main relevant principle for authorship is that authors must be able to take responsibility for their work. As a corollary, the listed authors must all be real people, not content creation tools such as large language models. Authors may use such tools in writing the paper, but their use must be fully disclosed (for example in the acknowledgements section at the end of the paper), specifying which tools were used, what tasks they were used for, and where in the paper they were used.
Reviewers must maintain higher standards than those required of authors. In particular, reviewers should not use generative AI tools to draft reviews. They may use such tools to explore questions raised by the paper, to search for relevant literature, to generate examples, and so on; but they must write the review themselves. The reviewer must of course also take responsibility for the full content of the review.
Aside from exploring the content of a submitted paper, we expect reviewers not to make use deliberate use of AI tools at all in writing their review. The reviewers have a responsibility to maintain confidentiality of submitted papers, so it is unacceptable to upload any of the paper content, metadata, or results to any third-party system that cannot guarantee to maintain confidentiality. And in contrast to a paper submitted for publication, the grammatical polish of a review does not matter so much.