Research Foundations › Module 5: Integrity and ethics

Using AI tools honestly

Lesson 3 of 3 · 2 min

Previous

AI tools like chat assistants can help with research: explaining a concept, fixing code, improving the grammar of a draft. They also create new ways to make mistakes, or to cheat.

AI cannot be an author

The Committee on Publication Ethics (COPE), whose guidance thousands of journals follow, states that AI tools cannot be listed as authors. An author must take responsibility for the work, approve the final version and answer for it, and a tool cannot do that. Major publishers, including Nature, say the same.

Say what you used

Most journals now ask authors to disclose how they used AI tools, usually in the methods or acknowledgements section. A simple, honest sentence is enough: "We used [tool] to improve the language of the introduction. All content was checked by the authors."

Check everything

AI tools can produce text that sounds right but is wrong. A known problem is made-up references: citations to papers that do not exist, or that say something different. Never cite a paper you have not opened and read yourself.

Know the line

Asking for an explanation of a concept you then check
Not fine: Pasting AI-written text as your own analysis or results
Help with grammar and wording
Not fine: Citing references you have not read
Help debugging your own code
Not fine: Generating or "improving" data or images
Brainstorming search keywords
Not fine: Uploading confidential or personal data to a public tool

Uploading unpublished work, private data or a paper you are reviewing to a public AI tool may break confidentiality. Check your institution's rules.

RememberYou are responsible for every word, number and reference in your work, whoever or whatever helped you write it.