Scientists say ‘artificial intelligence is here to stay’

Despite widespread distrust, a survey shows that the majority of researchers are currently using tools powered by artificial intelligence (AI). Earlier this year, the Scientific Advice Mechanism (SAM), the advisory body to the Commission, released its independent policy recommendations on the uptake of AI in EU research and innovation. Its experts believe using AI in science will bring both advantages and challenges.

It is an interesting paradox: although the vast majority of researchers in Europe are using some form of AI tools, they still have concerns about data protection, privacy, effectiveness and intellectual property rights when using AI. For many, AI technology with its algorithms is a ‘black box’ in terms of how this technology actually works.

Researchers admit that AI plays an ever-growing role in their daily work, according to a survey by Oxford University Press. AI accelerates discoveries and facilitates their work by mining data and summarising, analysing and presenting results. The use of AI in science serves many purposes, including large-scale data processing, generating patterns and predictions, designing and controlling experiments, as well as writing and peer-reviewing papers. “AI is here to stay”, the respondents believe.

In 2023, there were already over 1 000 EU-funded research projects related to AI, according to a foresight survey by the European Research Council (ERC). Many of the surveyed respondents expressed the view that using AI in science would require implementing ethical guidelines and standards to prevent misuse and misinformation.

The SAM, the independent scientific advice network to the European Commission, recommends developing standards to ensure the required transparency and trustworthiness of the data used when taking a human-centric approach to AI.

Researchers’ opinions

When the ERC survey asked researchers about how they currently use AI in their scientific practice, they highlighted AI’s supporting role. Many see AI as a co-pilot, in particular when it comes to tasks such as writing and editing.

“AI will blur boundaries between quantitative and qualitative research and make ‘mixed methods’ (…) the normal procedure. It will not replace humans in qualitative interpretation, but it will provide for a back-and-forth between methodological approaches that still today is extremely labour intensive” comments Eeva Luhtakallio, a survey respondent from the University of Helsinki, Finland.

In general, AI-based tools are currently used to extract information from documents, create images and restructure and adapt texts for different purposes. However, almost all of these tasks depend on the trustworthiness of the AI-based tools used, the quality of the output and the non-fraudulent use of data.

Professor Nicole Grobert, Chair of the Group of Chief Scientific Advisors (GCSA), part of the Scientific Advice Mechanism: “One priority for all scientists is the quality of data. For example, if you feed AI machines with ‘rubbish’ they will spit out ‘rubbish’. The output of the machines is only as good as the source they are trained with. No rubbish in – no rubbish out.” – Read the full interview with Nicole Grobert

Establish an AI institute

In its joint scientific opinion of the SAM, the GCSA called for a publicly funded AI institute to be set up that would provide resources, cloud infrastructure and AI training for scientists.

For the scientists contacted for the ERC foresight survey mentioned above, the most worrying aspect of AI is still the lack of transparency and replicability.

Some thinkers and researchers compare the uptake of AI to the Industrial Revolution, including 2024 Nobel Prize winner Geoffrey Hinton, a British-Canadian computer scientist and cognitive psychologist. Hinton acknowledges that AI could transform parts of society for the better, while also expressing concern that the very technology he helped create may eventually surpass human intelligence in unpredictable ways: “I am worried that the overall consequence of this might be systems more intelligent than us that eventually take control. We have no experience of what it’s like to have things smarter than us”, Hinton says.

More than 1 000 tech leaders and researchers, among others, have signed an open letter calling for a moratorium on the development of the most powerful AI systems.

To ensure human control over AI’s use of data, Professor Andrea Emilio Rizzoli, Director of the Dalle Molle Institute for Artificial Intelligence in Switzerland and co-chair of the SAM working group on the uptake of AI, recommends as a rule: “Keep the control on the processing pipeline and decide how to proceed at critical points.”
Read the full interview of Andrea Emilio Rizzoli

As for whether machines will soon make progress in creativity, Professor Rizzoli thinks that “we should not be too worried” when it comes to creativity. According to his observations, machines eat up what they are fed and then they spit it out in various statistical recombinations, but these machines still lack the edge needed for interpersonal communication and creativity.

Useful links:
• European Research Council Executive Agency: ‘Use and impact of artificial intelligence in the scientific process – foresight’, Publications Office of the European Union, 2023.
• Scientific Advice Mechanism to the European Commission: ‘Successful and timely uptake of artificial intelligence in science in the EU’, 2024