In the midst of the expansion of artificial intelligence, more and more AI models are trained with scientific data to infer answers to scientific questions. But, Can AI replace scientists? Alessandra Buccella, associate professor of Philosophy at the State University of New York, has contributed her point of view in an interesting article explaining the exclusively human aspects of the research.
The Trump administration announced at the end of the year the ‘Genesis Mission’, an initiative to build and train a series of AI agents with federal scientific data sets with the aim of «test new hypotheses, automate research workflows and accelerate scientific advances». So far, the results have been mixed. Although AI systems can process huge sets of data, his lack of sound reasoning may result in unrealistic or irrelevant experimental recommendations.
AI in Science
Buccella, as a philosopher who studies both the history and conceptual foundations of science, sees several problems with the idea that AI systems can “do science” without humans or even better than them. “While AI can help with tasks that are part of the scientific process, it is still far from automating science, and may never achieve it”he assures.
And AI models do not learn directly from the real world: their human designers must “inform” them about what the world is like. Without human scientists overseeing the construction of the digital world in which the model operates, the advances that AI facilitates would not be possible. To be successful as scientific tools, AI models must maintain a strong empirical link to already established knowledge. That is, the predictions a model makes must be based on what researchers already know about the natural world.
The role of scientists in the discovery and experimentation process goes beyond ensuring that AI models are properly designed and anchored to existing scientific knowledge. Science, as a creative achievement, derives its legitimacy from human capabilities, values and ways of life. These, in turn, are founded on the unique ways that human beings think, feel and act, and scientific discoveries are more than just theories backed by evidence: they are the product of generations of scientists with diverse interests and perspectives.
As other philosophers of science have noted, scientists are more like a tribe than simply passive recipients of scientific information. Researchers do not accumulate scientific knowledge by recording facts, but rather create it through expert practice, debate, and consensual standards based on social and political values.
AI is not a scientist
The computational power It has been key in human research in the last century and has accelerated scientific progress. AI can help even more. Well-designed and rigorously trained AI tools would facilitate, and even speed up, the more mechanistic parts of scientific research. These tools would collect information about what has been done in the past to facilitate the design of future experiments, the collection of measurements, and the formulation of theories.
But not further… “If the vision guiding the implementation of AI models in science is to replace human scientists or completely automate the scientific process, I believe the project would only turn science into a caricature of itself.”assures the philosopher. «The very existence of science as a source of reliable knowledge about the natural world depends fundamentally on human life: shared objectives, experiences and aspirations»he assures.
