AI-powered large language models like ChatGPT, Grammarly, and various other writing assistants have become a staple for scientists and students alike. Whether it’s to help you refine ideas, restructure sentences, or proofread for clarity, there’s no doubt that AI is playing an increasingly prominent role in academic writing.

But here’s the question many researchers ask me as an editor and proofreader: Do I need to disclose that I used AI?

The answer? It depends—and that’s where things get tricky.

If you’ve tried to find a clear answer, you’re not alone. Guidelines on AI use vary significantly between universities, journals and publishers. Some have explicit rules, while others remain vague, which leaves researchers to interpret policies on their own.

The main concerns revolve around:

For many researchers, this lack of consistency adds another layer of stress to an already complex publishing process.

Rather than leaving you to navigate this on your own, we did the work for you. We compiled a number of AI-related policies from various universities, journals, and publishers to provide some clarity.

Here’s what we found:

So, what should you do if you’re using AI in your research writing? Here’s the best approach based on current guidelines:

Right now, AI policies are clear on one thing: generating large chunks of content with AI is a no-go. However, the use of AI for proofreading and editing is still a gray area.

Until publishers offer clearer guidelines, the safest bet is to check policies carefully and use AI in a way that enhances—not replaces—your academic work.