Before incorporating generative AI into your research activity (bibliographical review, writing, analysis), it is necessary to fully understand its capabilities and, above all, its limitations.

How can you do this?Take the time to understand how generative AI models work, to gain an understanding of what they can do and to avoid avoid dependence on them. Make sure your use of AI is transparent and does not compromise either integrity or scientific credibility.

Generative AI not only makes mistakes but also tends to invent content such as references, oversimplify complex concepts and perpetuate biases present in the training data.

What can you do?Treat any text, bibliographical reference or idea generated by AI as a starting point, and verify the results and each reference using academic search engines (Google Scholar, PubMed, Scopus). Question the results based on your expert knowledge of the field. For reproducible science, it is advisable to explicitly document and declare the use of generative AI.

Do not share sensitive, confidential data, data subject to material transfer agreements or intellectual property rights with generative AI applications from third parties.

How can you do this?Before using a generative AI tool to process research data (especially if it involves personal data from human subjects or unpublished results), ensure that the platform complies with data protection regulations (GDPR) and that the servers offer confidentiality guarantees. Remember that, when introducing an idea or a result, you may be relinquishing ownership of it or revealing it prematurely.

Generative AI can be useful for drafting, debugging code or suggesting ideas, but the ultimate responsibility for the integrity, originality and validity of the research lies solely with you.

How can you do this?Never include generative AI as the author of an article, as indicated by the main editorial guides. Ensure that the intellectual core (hypothesis, experimental design, data interpretation, conclusions) is based on your own expertise. And critically assess how the use of generative AI could influence the direction of your research to prevent unwanted bias.