If a reviewer asked you to specify exactly how AI was used in your research project, would you have a clear answer?

Researchers are already using AI across the research process but expectations are shifting. Institutions, journals and funders are increasingly asking not just IF AI was used but HOW it was used.

Researchers are being asked to specify what AI tool was used and how, at what stage of the research process and the extent to which outputs were evaluated or revised. However, there is little guidance on consistent reporting. Without a framework, disclosures can be vague or inconsistent.

That gap is what led to the development of the AIR Framework. ⚙️

AIR is a structured framework to help researchers make AI use explicit, transparent and defensible. It does not tell you WHETHER or not to use AI. It helps you describe how and where you are using/ intend to use it, what level of judgement is required and what risks may follow.

If you are leading a research group, supporting doctoral researchers or thinking about how to document AI use more clearly, this may be useful.

You can get the AIR framework here:

**********

AIR: AI in Research – A framework for transparent and responsible AI use mapped to the research process

The AIR matrix is designed to help researchers describe, think through and explain how they use AI across the research process.

Download AIR & the how to guide from Figshare (link in the comments.)

AIR is shared under a Creative Commons licence CC BY 4.0 so anyone can freely use, share and adapt it, provided they give appropriate credit (attribution) to the original creator and link to the license.

**********

🙏 Thank you to all of the fantastic folks across researcher development, academia and publishing who gave us feedback on AIR during its development. Much appreciated and these discussions led to so many more potential follow up projects and ideas!

❓ How are you currently being asked to report AI use in your context and where is that guidance still unclear?”