Responsabilité Sociale des Universités
Formation Continue


AI to verify research reproducibility. The golden rule of research – that an experiment reproduced in similar conditions will yield similar results – has been under strain in the social sciences. A survey conducted by the Center for Open Science in Virginia shows that the findings in more than one in three major social science studies published in Science and Nature between 2010 and 2015 could not be replicated. The US government has therefore decided to fund a $7.6 million project aimed at using algorithms to create “confidence scores” for the reliability of social research papers. The success of the project, however, will depend on whether algorithms can provide evidence that is reliable and valid. Bid to use AI to predict research reproducibility launched, Times Higher Education, 14-20 February 2019.

China on a path to lead the world in scientific research? China’s rivalry with the US for global hegemony also goes through higher education, as evidenced by the huge investments the Chinese government has been making in scientific research. Spending on R & D grew tenfold between 2000 and 2016. In sheer numbers of articles, reviews and conference papers indexed in Elsevier’s Scopus database, China increased its production by a staggering 1,322% between 1997 and 2017, while the US output grew by 64% and the EU output by 111% during the same period. According to some estimates, China could surpass the US by 2022, and has already taken the lead in a number of fields such as computer science and engineering.

The quantity of papers published, however, is not the sole measure of success in scientific production. When it comes to China, the fact remains that the quality of some articles is very low, and wrong-doing or game-playing (plagiarism, suspected peer-review fraud) is always a concern – though Chinese researchers are increasingly publishing in the most prestigious scientific journals. The pressure on scientists to make breakthroughs and publish is such that ethical standards can easily be ignored, and intellectual property theft is widespread. This is causing the international scientific community to lose trust in Chinese research and view the prospect of collaboration with reluctance. The Chinese government is keenly aware of the reputational risk involved, and it announced last year a range of reforms aimed at clamping down on academic misconduct.

More broadly, China is very much counting on science to project its global power. To that end, it is developing a scientific establishment based on a core group of nine elite universities (known as the C9) and the Chinese Academy of Science, an official agency, that will abide by international standards. In 2008, China had already put in place the Thousand Talents program, which provided plenty of incentives for Chinese researchers working abroad to return to the homeland. Many did.

Today, China spends billions of dollars on dark matter and neutrinos detectors, is a world leader in applying quantum mechanics to computation and cryptography, has the world’s largest telescope, conducts world-class research in artificial intelligence, invests heavily in genome sequencing, gene editing, stem-cell research and new clean-energy sources, not to mention human space flight.

Scientific circles in the West view these successes with a measure of concern. In addition to the reluctance to collaborate, some observers, more broadly, fear that these huge investments and advances in STEM fields may have a “gravitational” effect on global research and pull it away even further from the humanities and social sciences.

On the whole, however, Chinese science thus far has been busy catching up rather leading the way. In addition, under Xi Jinping, the Communist Party is strengthening its stranglehold on institutions of higher learning, with far more interest in exercising control than in boosting creativity. Yet scientific inquiry requires free flows of information, the ability to doubt and apply critical thinking, question assumptions, share data with colleagues and even rivals, and, most importantly, speak truth to power. In particular, it thrives on collaboration and exchange. Neither the Party not the political culture can at this moment encourage such practices.

In the long run, this may place a permanent limitation on the capacity of China’s universities to lead.

Red moon rising, The Economist, January 12th 2019. The great experiment, The Economist, January 12th 2019. A new centre of gravity, Times Higher Education, 14-20 February 2019.

Auteurs : Corinne Mellul
Date du document : avril 2019
Prospectives : La recherche demain
Thèmes : Autres