Salomon is Dean and Professor of Decision Science at Amsterdam Business School. Boot is Professor of Corporate Finance and Financial Markets at the same school. They argue that university strategies often focus on developing AI, using it in education and maintaining digital independence. Yet they pay too little attention to AI’s implications for the research process and for the people conducting research.

Recent developments illustrate the speed of change. AI has produced new results in difficult mathematical problems and is increasingly able to carry out research tasks that previously required substantial human effort. According to Salomon and Boot, this does not make researchers redundant. Instead, their role will change.

Evaluating AI’s work

Researchers will increasingly need to design and guide AI systems, and to assess their output critically. They must determine whether findings are reliable, original, relevant and open to verification. This is particularly important as AI can rapidly produce large amounts of research that may appear convincing without having been sufficiently checked.

The authors also warn against reducing investment in junior researchers. AI may make some routine research work more efficient. However, reducing early-career positions could weaken the future pipeline of researchers who have the methodological knowledge and subject expertise to evaluate AI-generated work.

Finally, they stress that AI will affect disciplines differently. In mathematics, many steps in a proof can be formally checked. In fields such as economics, business administration and sociology, findings often depend more strongly on interpretation, context and professional judgement.

Scientific integrity and quality

Salomon and Boot therefore call on universities to formulate research-specific AI strategies for individual disciplines. Such strategies should safeguard the human expertise needed to maintain the quality and integrity of science.