Speaker
Description
The world of scientific research has been in crisis for two decades due to the ‘publish or perish’ culture, which has led some researchers to cheat on their results or resort to paper mills to get published. Two aggravating factors have been recently added: Covid crisis, which has led the general public to doubt the work of researchers, and the development of LLMs, which can create a scientific article in a matter of minutes.
We will first see that generative AI has created an environment that enables fraud: detection software cannot tell whether AI was used as a tool to improve the quality of writing or to produce an entire article. We will list the means available for detecting fraud.
For the second part, which concerns the Covid crisis, we will focus on France: By establishing himself as a world authority (with 3,500 publications) and proposing a simple solution to the pandemic (the use of hydroxychloroquine), Didier Raoult became a celebrity. Around this figure, a group of scientists known as ‘rassuristes’, emerged, explaining that Covid was not dangerous, that there would be no second wave, and that vaccines were useless. We conduct a semiotic analysis of the discourse of a group of ‘rassuristes’, the CSI, from 2020 to 2025. Beneath a classic academic appearance, the statements are problematic. Anti-vaccine rhetoric, for example, is no longer directed against science in general, but against ‘official’ science (which includes the WHO, Bill Gates, etc.), with CSI representing the good side of science. There is clearly a problem of ultracrepidarianism: they are genuine researchers, but what expertise do they have in vaccines when they are mathematicians or computer scientists?
We conclude that ‘rassuristes’ reinforce traditional anti-scientific rhetoric by giving their readers the impression that they are on the right track, the path to a hidden truth that will eventually be revealed. The solution to the research crisis lies in a return to traditional methods.
| Keywords | science crisis, scientific integrity, ultracrepidarianism, AI, trust in science |
|---|---|
| jean-louis.vaxelaire@unamur.be |