AI Has a Darker Side in Healthcare, Expert Says

Los Angeles -- AI possesses advantages in healthcare, yet experts warn of potential drawbacks, according to a senior researcher at the Technology and Work Program of the UC Berkeley Labor Center. Mishal Khan, PhD, highlighted that even if workers do not lose their jobs due to AI, there can still be negative impacts on their well-being.

Khan classified AI's applications in the workplace into three categories: electronic monitoring, algorithmic management, and task or job automation. Electronic monitoring involves using cameras to observe workers for signs of fatigue or to ensure they're using safety equipment, while algorithmic management may include AI listening in on phone conversations between workers and customers and coaching workers on how to respond.

Task automation refers to specific parts of a job being outsourced to AI. Despite AI's potential benefits, its widespread use could lead to job loss, theft of work product, job de-skilling, lower wages, work speed-up and injuries, discrimination, suppression of labor organizing, and loss of autonomy and control, according to Khan.

In her research on AI's impact on healthcare, Khan found that de-skilling, where humans lose specialized skills due to AI or other automation, is a significant concern and could threaten healthcare workers' scope of practice. Tina Hernandez-Boussard, PhD, a professor of biomedical informatics at Stanford University, echoed Khan's sentiments, noting that medical residents who went to Uganda to assist at a cardiac ward were unable to confidently recommend treatments due to a lack of electricity and AI.

Aliénation and detachment from patients could also result from automated patient handoffs, and liability issues arise if mistakes are made as a result of AI recommendations. Additionally, Khan expressed concerns about the spread of ambient scribes and gigification of healthcare jobs, as well as state legislations addressing some of these issues.