Researchers hold a wide range of views on AI. Some use it every day, some are experimenting, and some prefer to wait and watch. This post is not here to pick a side. Whatever your view, the goal is the same: work that other people can read, check and trust. So let us look at what that takes in a world where research is getting faster.

What has not changed

Credibility has always rested on a few simple things: methods that others can follow, data that others can check, honest reporting of what was found, and authors who stand behind their work. None of that depends on which tools were used along the way. A careful study done with a pen and paper and a careful study done with new software are trusted for the same reasons.

What has changed

The pace. Drafts, searches and analyses that once took days can now take hours, and more research is being produced as a result. Many researchers find AI tools useful for scanning a large body of papers, tidying the language of a draft, explaining a method in plain words or helping with code. Others prefer to do these things by hand. Both are fine.

What is new is a shared question: how do we make sure that faster also means sound? Journals are updating their policies, reviewers are seeing more submissions, and every researcher is working out which habits suit their own work.

What journals and ethics bodies say

Most major publishers and bodies such as COPE and the ICMJE give broadly similar guidance. The authors remain fully responsible for everything in a paper, whichever tools helped to produce it. An AI tool is not listed as an author. And where a journal asks for it, the use of AI is disclosed. Rules differ from journal to journal and are still changing, so the best source is always the author guidelines of the journal you are writing for.

Good habits that work for everyone

These are useful whether you use AI heavily, lightly or never.

1. Check what you rely on. Open each reference, confirm that it exists and says what you claim, and reproduce the key numbers yourself, whoever or whatever first produced them.

2. Be open about your methods. A clear sentence on how a paper was made, including any tools, helps readers trust it.

3. Keep a simple record of your process. A short note of what you did and when makes it easy to explain your work later.

4. Protect confidential material. Check your institution's and the journal's rules before sharing unpublished data or someone else's manuscript with any tool or service.

5. Keep the thinking your own. The question you ask, the argument you build and the conclusions you draw are what your name stands for.

Credibility is a shared effort

Research becomes believable when people who know the field have looked at it. A co-author can question an assumption before a reviewer does, and a colleague at another institution can look at your data with fresh eyes. These human checks matter at every stage of work. If you would like more people like that around you, our guide on finding research collaborators outside your own institution is a good place to start.

The bottom line

Whether you welcome new tools or approach them with care, credible research comes from the same place: clear methods, honest reporting and good colleagues. There is room for every approach, and everyone benefits when the work can be trusted.