The strong connection between AI and open science is crucial for maximising the potential of scientific research. While AI represents rapid scientific advances, its integration must align with open science principles to ensure replicability, credibility, equity and trustworthiness. Open science offers mutual benefits for AI by expanding participation in research and promoting inclusive practices. The open science approach, when combined with AI technologies, can significantly contribute to addressing the current pressing global threats.
Associated with these opportunities are diverse challenges arising from the increased adoption of AI. These challenges such as the black-box nature of AI systems and the dominance of private companies in advancements pose significant obstacles. These include the reproducibility crisis (in which other researchers cannot replicate experiments conducted using AI tools); interdisciplinarity (where limited collaboration between AI and non-AI disciplines can lead to an inconsistent uptake of AI across science); and data challenges (while high quality data is foundational to AI applications, researchers consistently face challenges related to its volume, heterogeneity and potential for bias). There are fundamental barriers limiting the adoption of open science principles, due to the black-box nature of AI systems and the phenomenon of private companies outpacing academic institutions at AI-powered research. Furthermore, the changing incentives across the scientific ecosystem may be increasing pressure on researchers to be ‘good at AI’ rather than ‘good at science’.
Addressing these challenges requires applying open science principles to AI, particularly focusing on transparency. Overcoming cultural and economic barriers is essential to fostering a culture of open science in AI-driven scientific research.
