During the Covid-19 pandemic, the refusal to be vaccinated stemmed in particular from a lack of confidence in the vaccine’s efficacy and safety. Climate change deniers ( 32% in France ) continue to deny the role of human activities in climate change. These two examples share a common thread: the questioning of scientific findings, leading to “scientific skepticism . “

The integration of artificial intelligence at all levels of scientific production could reinforce this feeling among the public. Research that is ethical, responsible, and conducted with integrity is essential for maintaining public trust. Faced with the upheavals brought about by generative artificial intelligence (GAI) and in the absence of clearly disseminated rules within the scientific community, researchers in the humanities and social sciences are questioning the evolution and ethics of their practices.

In a highly competitive environment where researchers’ careers largely depend on their scientific output, using AI tools may appear as an option—at a supposedly minimal cost (most tools are free and seemingly easy to use)—to increase their own productivity by leveraging the capabilities of AI for synthesizing large datasets and writing. This delegation may seem appealing, but it carries significant risks.

A twofold question for researchers

The quality of scientific output relies on a peer – review system , most often anonymous. In the publication process, researchers are either authors or reviewers, with somewhat different perspectives.

• For the authors, AI can be used at various stages of the research process (literature review, data collection and analysis, writing, etc.), which raises questions about the very notion of intellectual property. In the humanities and social sciences, where the research process is often iterative, interpretive, and context-sensitive (due to its subject matter—society and human relationships—and for which qualitative methodologies are frequently employed; for example, an analysis of patient discourse during the adoption of an e-health solution), the use of AI can thus profoundly transform research practices. While it can assist researchers, it also raises a central question: at what point does this support become a delegation of the scientific act itself? However, it is a fundamental principle that, regardless of AI’s role in writing a scientific article, only the human author(s) will ultimately be responsible for the content produced. Therefore, it becomes urgent to develop shared, explicit guidelines on how to integrate it responsibly and declare its use transparently .
• For reviewers, the IAG (Integrated Analysis of General Principles) can be integrated into their own peer- review process . Indeed, reviewers may be tempted to use it to produce their evaluation (describing/summarizing the article, identifying its strengths and weaknesses, writing certain sections of their report). The IAG may also have been used by the article’s authors, and it is then up to the reviewers to detect this and assess whether its use respects the scientific integrity required by the journal.

The use of the Problematic Paper Screener (PPS) tool, created by Guillaume Cabanac, has made it possible to detect, to date, 25,000 articles using “tortured expressions” suggesting the use of an algorithm for their writing, without discernment or human review.

However, although many journals have now established rules, ethical principles remain vague and practices are heterogeneous, leaving researchers in uncertainty.

Also, it is clear from the speeches and debates of learned societies that there is a real need to dialogue on these issues , but also a lack of unanimity on the options available, with divergent ethical perceptions .

Towards an impoverishment of thought?

A recent study published in the journal Science shows that the adoption of large language models (LLMs) by researchers (across all disciplines) increases their productivity (more than 89%) and opens up new opportunities (simulation of virtual populations when these are difficult to access; rapid synthesis of a collection of research articles) but that it also reduces the quality of their work due to its limited capacity to grasp complex tasks.

Furthermore, relying on AI to acquire knowledge and write carries a risk of standardizing thought and knowledge: its use tends to confine researchers to well-established fields, thereby reducing scientific exploration “outside the box.” Researchers aware of these limitations can benefit from using AI by querying it precisely and thoroughly, not settling for the first answer, and by systematically verifying the content produced.

The use of IAG for article evaluation: a major risk to the reliability of research and its usefulness to society?

Intelligent machine translation (IAT) presents a major problem because the texts it generates are currently impossible to automatically distinguish from those produced by humans. Therefore, at this stage, two principles must be implemented to overcome this limitation.

The researcher must be the sole guarantor of the scientific quality of the research: researchers are asked to evaluate articles precisely because they are experts in the scientific field in question. Their analysis should therefore theoretically stem from their own judgment, without delegation to an IAG (Integrated Assessment Tool). This is also specified in the ethical charters or guidelines of some academic journal publishers (such as Elsevier or Sage ), which, at this stage, prohibit reviewers from uploading an article to an IAG tool. However, if the IAG becomes a support tool, what limits should be placed on its use?

The confidentiality of research data is becoming a guarantee of the quality of future research: using an AI hosted in the cloud to evaluate an article risks exposing confidential and potentially sensitive data to everyone, which is particularly problematic in the context of evaluating a scientific article. It could be considered to use an AI installed locally, not only to preserve the confidentiality of research data but also to avoid feeding the AI ​​with scientifically unvalidated documents. The credibility of research results and the public’s trust in scientific discourse depend on it.

The rapid development of AI in the writing and evaluation of scientific articles necessitates, on the one hand, a rethinking of our relationship to knowledge and its construction, and on the other hand, a profound reflection on how to preserve the ethical foundations of knowledge production and evaluation. Despite some promising initial initiatives , a collective, societal, interdisciplinary, and international reflection appears essential to reconcile technological innovation with scientific responsibility.

Author Bios: Nathalie Guichard is a Professor at Paris-Saclay University, Agnès Helme-Guizon is Professor of Social Marketing, Grenoble IAE Graduate School of Management at Université Grenoble Alpes (UGA), Anne-Sophie Cases is Professor, MRM Laboratory at the University of Montpellier, Christelle Aubert Hassouni is a Lecturer and researcher, specialist in consumer privacy issues at ESCP Business School and Jean-François Toti is a Senior Lecturer in Management Sciences – Marketing at the University of Lille