Mishal Khan, a senior researcher at the UC Berkeley Labor Center, highlighted potential pitfalls of artificial intelligence within the healthcare sector. She categorized AI’s uses into electronic monitoring, algorithmic management, and task automation. Electronic monitoring involves the use of cameras to observe workers, such as in transportation or construction. Algorithmic management utilizes AI software to analyze conversations and provide coaching to workers. Task automation involves outsourcing specific job functions to AI. Khan stated that in the absence of real power, the use of AI risks developing or exacerbating potential harms.
These risks include job loss, theft of work product, job de-skilling, lower wages, discrimination, suppression of labor organizing, loss of autonomy, and alienation from patients. Khan’s research identified a concern regarding de-skilling, particularly threatening healthcare workers’ scope of practice. Tina Hernandez-Boussard, a professor at Stanford University, noted an instance where medical residents lacked AI support during an outage, hindering their ability to synthesize information. Concerns were raised regarding data collection by ambient scribes and the potential for algorithmic management to drive the ‘gigification’ of healthcare jobs. State legislatures are enacting laws to address these issues, including prohibitions on replacing mental health workers with AI and preventing the training of AI replacements.”}