Source: French to English Tester Published on: 2026-05-04
Source: The Conversation – France (in French)– By Dejan Glavas, Associate Professor of Finance, ESSCA School of Management, National School of Bridges and Roads (ENPC)
The impact of generative artificial intelligence on social sciences can be understood through the metaphor of the novelThings Fall Apartby Chinua Achebe. This book traces the upheavals of Nigerian society in the face of colonial intrusion. It offers a relevant approach to understanding current tensions in the academic world.
A few months ago now, a colleague sent me his article for a second reading. Everything was there: a finely honed problem statement, an exhaustive literature review, rigorous methodology. Too much, actually. Upon closer inspection, some expressions alerted me: “crucial,” “in the dynamic world of,” “it is imperative to note that.” The typical vocabulary of ChatGPT. I asked for explanations. Awkward response: “Yes, I used AI to write certain parts, but I obviously reread it.”
This discomfort reveals an unsettling truth: we, social science researchers, are losing control of our profession. And we are pretending not to see it. This reflection led me to propose aarticle to theFrench Management ReviewOn the occasion of its 50th anniversary issue.
Some illustrative figures of the paradigm shift
About 13.5% of biomedical article abstracts published in 2024 contain traces of the use of generative artificial intelligence (AI) according to astudypublished in Science Advances. Even more worrying: astudyStanford researchers reveal that up to 17% of the text in article reviews at major AI conferences may have been substantially modified by AI itself. We have entered a loop where AI evaluates AI.
Furthermore, research teams analyze corpora in a few weeks that would have required years of work. In digital sociology, for example, astudypublished inScienceexamined the spread of nearly 126,000 true and false pieces of information shared by approximately three million users on Twitter, amounting to more than 4.5 million shares, an unimaginable volume without automated analysis tools.
Breakdowns in social sciences
But beyond these productivity gains, something is cracking. AI does not just disrupt our ways of working. It calls into question what it means to “do research.”
How to validate results produced by an algorithm whose internal workings remain opaque? How to ensure scientific rigor when AI can “hallucinate” and invent bibliographic references that do not exist? How to protect our critical thinking in this context?
Also to read:
Management sciences: a medicine to heal organizations
We are currently facing the same dilemma as the characters ofnovelof the Nigerian writer Chinua Achebe, entitledEverything is collapsing(Things Fall Apart, in English). In this story, a traditional African society faces European missionaries who come to impose new beliefs. Some members of this traditional society adapt to this arrival, others resist. Everyone understands that nothing will be the same as before.
Technophiles against traditionalists
The scientific community is currently divided. Some, the “technophiles,” welcome AI with enthusiasm, convinced that it will accelerate discoveries and open research to the developing world. Others, the “traditionalists,” denounce a race for productivity at the expense of depth of thought.
This fracture is not limited to technical and technological issues; it also affects our identity, both collective and individual, as researchers. Are we becoming mere operators validating productions generated by algorithms? Is our expertise reduced to knowing how to ask the right questions to an AI model?
Young researchers are the most exposed. They juggle between the pressure to publish quickly and the implicit injunction not to use AI too much. Result: many use it secretly, creating a culture of secrecy that has consequences for scientific transparency.
What solutions?
We have preferred to look away for too long. Artificial intelligence is here, and ignoring its presence is no longer tenable. We must rethink our way of doing science. Imagine a new framework.
Three main principles could serve as our compass.
The first is transparency. Every use of AI should be specified in black and white in our publications: which tools were used, for what purposes, and within which limits. Some journals, likeNature, are already starting to impose themselves.
Next, responsibility. AI can analyze data, suggest leads, draft outlines. But it is up to the researcher to interpret and validate with a critical mind. This red line must never be crossed. OnestudyA recent study cruelly reminds us: 52% of answers generated by AI to programming questions contain errors, and human supervision fails to correct them in 39% of cases.
Finally, pluralism. It becomes necessary to accept that some researchers integrate AI into their work while others reject it. This diversity can become a strength if debates take place openly, in conferences and seminars that constitute the scientific tradition.
A new scientific paradigm
Artificial intelligence will continue to gain in power and ubiquity.
We are on the brink of a turning point in the world of research where human intelligence becomes one of the forms of intelligence. It seems clear that this coexistence will only bear fruit if it is based on rules established by the researchers themselves.
The real risk does not come from artificial intelligence, but from our inertia. Our desire to take advantage of what it offers without questioning too much. To accept, without resisting, an ease that could, in the long run, slip away from us.Chinua Achebe narrateshow a society collapses when it loses control of its destiny. We, researchers, still have a choice. We can decide how to integrate artificial intelligence, under what conditions, but also with what limits.
The challenge is not to resist change. It is to guide it towards a more rigorous, more transparent, more self-aware science. A science that uses AI without submitting to it. A science that remains, above all, a human adventure.
Dejan Glavas does not work for, advise, hold shares in, receive funds from any organization that could benefit from this article, and has declared no other affiliation than his research institute.
–ref. Will artificial intelligence put an end to research in social sciences?https://theconversation.com/will-artificial-intelligence-end-research-in-social-sciences-281380