There must urgently be guidelines for dealing with AI in science, says the Rathenau Institute. Incorrect use ‘undermines the reliability of scientific knowledge and trust in science’.
The call is a response to a report in the Dutch newspaper Volkskrant, which describes how AI is advancing science in various disciplines.
But there are many risks associated with AI, the Rathenau Institute warns. Generative AI, such as ChatGPT, puts ‘core scientific values under pressure’, such as integrity, independence and reliability. The Rathenau Institute also sees a strong rise in AI-generated fake articles. Clear rules must therefore be introduced quickly.
Tracing research
According to a survey by scientific publisher Elsevier at the end of last year, around 58 per cent of scientists now use (in Dutch) some form of artificial intelligence. It is mainly used to trace recent research and to gather literature.
The Rathenau Institute calls on universities and other knowledge institutions to establish clear guidelines. It also raises the fundamental question: ‘what kind of science do we actually want, with or without generative AI?’
Code of conduct
Meanwhile, the Royal Netherlands Academy of Arts and Sciences (KNAW) is currently working on a new code of conduct for scientific integrity. A draft version was released last autumn. Artificial intelligence is also addressed in it.
“Use only technologies whose functionality is known and scientifically validated,” the draft code states. This is difficult with AI programmes such as ChatGPT and Perplexity, where it is often unclear which sources they are based on and how data are processed.
The new code of conduct will apply to both universities and universities of applied sciences and is due to come into force in the autumn, according to the KNAW.
HOP, Naomi Bergshoeff
Comments are closed.