Routinely handing work tasks over to neural networks is quietly weakening specialists' expertise and making it harder for newcomers to learn their profession, according to Kaisa Kärki, a researcher at the University of Helsinki who studied the phenomenon of deskilling — when people lose professional skills, or never manage to acquire them, because artificial intelligence does the work for them, infohub.kz reports.
Medicine served as the main example in the study. According to the author, human skills weaken without constant practice, yet specialists do not notice this when working with a neural network, since the task is formally completed.
After artificial intelligence was introduced, for instance, endoscopists working without prompts from a neural network detected precancerous polyps in 22.4% of patients, down from 28.4% previously. A similar decline in competence was recorded even among experienced accountants with more than ten years on the job.
Kärki argues that professional communities, rather than individual workers, should take responsibility for preserving core knowledge. She proposes mandatory periods of work without neural networks and requiring specialists to make decisions before consulting artificial intelligence — by analogy with pilots, who regularly train to fly without autopilot.