Ad: BlueJ Better Tax Answers. -Accomplish hours of research in seconds -Instantly draft high-quality communications -Verify answers using a library of trusted tax content. Learn more

Scholastica Announces AI Disclosures for Law Review Submissions

In November, Scholastica, the (currently) leading portal for submissions to law reviews, announced an opt-in framework for AI disclosures for article submissions. I want to be clear: these disclosures are necessary. Essential! But, oh wow, I’m totally not sure Scholastica hit the mark.

And, given law review editors’ relative inexperience and high turnover, the consequences of a major whiff are bigger in law than other academic fields. Scholastica’s released standards distinctly show a lack of specificity, a tendency towards overbreadth, and an absence of reciprocal AI-disclosure standards for editorial teams.

Scholastica’s Common Guidelines and some commentary, below the fold.

Can law reviews require AI disclosures for Scholastica pool submissions?

Starting in 2027 — yes!

Exclamation point!

The Common Law Review Generative AI Disclosure Guidelines

Law reviews that have adopted these common guidelines require authors to disclose the use of generative AI in the preparation of their submission for the following activities:

  • Developing manuscript ideas or arguments, including researching or synthesizing information, brainstorming, framing, and analysis
  • Substantively writing, arranging, or editing any portion of the submission, excluding basic spelling, grammar, and copyediting assistance
  • Generating, formatting, or summarizing citations, quotations, or parentheticals
  • Processing or analyzing empirical datasets, translating foreign-language sources, or creating or modifying figures, charts, or images

For each required disclosure in the above categories, you must input the AI tool(s) you used, the version of each tool if known, an explanation of what you used each tool for, and how you verified and built on the output.

Okay, okay. So what, exactly, is “generative AI”? Consider public schools’ recent struggles with unbundling AI features from Google search. Scholastica gives participating law reviews “the option to require authors to provide the AI prompts, inputs, and outputs they used.” Is that every Google search? Any query that CoCounsel touches? Or just the stuff authors run by their buddy Chat? Legal scholarship shouldn’t lag the physical and social sciences in defining its terms. Confidentiality issues also play into these disclosure policies, which almost certainly shouldn’t mirror discovery in federal courts.

And how about “Substantively writing”? One issue is intellectual property. The Copyright Office takes the position that purely AI-generated text is not protected, while sufficiently human-generated text is, with a lot of grey where the two intersect. If an author checks box number two, how does the journal assert its or the authors’ copyright in the published product? Should protections be established on a sentence level or more holistically? And are today’s author agreements capacious enough to cover these nuances?

Lawyers know that naming an issue sometimes raises all of the other issues. That’s one reason that the billable hour has staying power. Other publishers’ disclosure schemes may avoid this intellectual property issue by requiring a first human draft of any ultimately published material. That’s not an approach that’s really contemplated by Scholastica’s guidelines, which seem likely to yield either amorphous disclosure requirements or blanket prohibitions on AI use. Neither extreme is a likely settling point for academic norms.

Finally, Scholastica’s guidance advises (in bold, but not boldly) that student editors should “Keep in mind that AI use policies should go both ways.” Editors are exhorted “to communicate whether your law review allows editors to use AI tools to support article review and, if so, how, so authors are aware of that.” Great and well-said.

But, (1) AI tools probably should be used (responsibly) in the editorial process; that’s clearly where publication will go over time. Also, (2) the biggest standardization gap is on the publication end, not authors’ side. A menu of standard AI agents or a recommended workflow could improve the editorial process significantly, and Scholastica could “fill this gap,” as it were, through the platform’s (current) market position in the field of law.

Across academic fields, the process of developing AI disclosures for manuscript submissions will be iterative. Half of that’s the fact that AI (and AI use) are changing rapidly, leaving the lines between “work” and “AI-assisted work” increasingly unclear. But the other half reflects, I think, a sense that getting something out the door is better than waiting and seeing. Maybe Scholastica’s AI-disclosure policy is the right move, but its infirmities don’t sit well in the context of the idiosyncratic publication process for legal scholarship or Scholastica’s power within that market.

Related TaxProf Blog coverage:


About the Author

Ad: BlueJ Better Tax Answers. Blue J's generative AI tax research solution is transforming how tax experts work. Learn more.
Information and rates on advertising on TaxProf Blog