A new study from the Institute for Communication Science at RWTH Aachen University found that experts are significantly more likely to rate AI developments as beneficial and less risky compared to the general public. This finding was reported by the university on Friday.
The research surveyed 1,100 citizens and 119 AI experts. The study examined various AI scenarios, ranging from medical diagnosis and autonomous weaponry to political decision-making. Study leader Philipp Brauner noted that among experts, the overall assessment of AI is overwhelmingly determined by perceived benefits rather than risk. He observed that this focus on utility is much less pronounced among the general population, where risks weighed more heavily.
The researchers raised concern about a structural risk: if AI is developed primarily with a focus on utility, the resulting systems may overlook the risk priorities of the broader public. They term this potential pitfall “Procrustean AI”. To counter this, the study advocates for greater public involvement in the development, implementation, and regulation of AI. This research was published in the journal “AI & Society”.