AI Has A Darker Side In Healthcare, Expert Says

TL;DR

For listenersOffer from Amazon

Turn quiet afternoons into listening time

• Thousands of audiobooks, podcasts and originals
• Listen on your phone, tablet or Echo — also offline
• Cancel anytime

Try Audible free →Free trial for new members

As an affiliate, we earn on qualifying purchases.Hide

Labor researcher Mishal Khan warned at the Association of Health Care Journalists Fall Summit that AI in healthcare can harm workers even when it does not replace their jobs. She pointed to risks including loss of skills and autonomy, surveillance and changes to patient relationships; the scale of these effects remains unclear.

Mishal Khan, a senior researcher at the UC Berkeley Labor Center, warned that artificial intelligence could worsen working conditions in healthcare even if it does not eliminate jobs. Speaking at the Association of Health Care Journalists Fall Summit in Los Angeles on October 13, she identified concerns including worker surveillance, loss of specialized skills and reduced control over work.

Khan described three broad ways AI is used in workplaces: electronic monitoring, algorithmic management and automation of particular tasks. In healthcare, these uses could affect how work is tracked, how workers interact with patients and how decisions are made about which tasks people perform. The report did not identify a specific healthcare employer or system as the subject of the warning.

In research on AI’s possible effects in healthcare, Khan identified de-skilling as a particular concern. If software takes over specialized tasks, workers may have fewer chances to practice them, potentially affecting their skills and scope of practice. Stanford biomedical informatics professor Tina Hernandez-Boussard described residents who, after traveling to a hospital in Uganda, faced a loss of electricity and had no AI system to read electrocardiograms. She said the residents were not confident synthesizing the information to make recommendations.

Khan also raised concerns about automated patient handoffs and ambient documentation tools, often called ambient scribes. Workers, she said, are asking what information these systems collect about them and whether it could be used in employment decisions. Other risks she listed included job losses, lower wages, work speed-up, injuries, discrimination and limits on labor organizing. These were presented as potential harms, not as outcomes shown to be occurring across healthcare.

AI Could Change Clinical Work

The concerns matter because clinical work depends on both technical knowledge and workers’ ability to exercise judgment. If automation reduces opportunities to practice tasks, clinicians could be less prepared when systems are unavailable, as Hernandez-Boussard’s example illustrates. Her account does not establish how common such a problem is, but it points to a possible consequence of relying on tools without maintaining human expertise.

AI can also shift power at work. Monitoring and algorithmic management may give employers more information about employees and influence how tasks or shifts are assigned. Khan warned that algorithmically managed nursing platforms could require nurses to bid for shifts, with algorithms potentially estimating the lowest wage a nurse might accept. That was a projected scenario, not a report that such a system is already in widespread use.

For patients, the issues include continuity of care and accountability. Khan said workers are concerned about the effects of automated handoffs on relationships with patients. She also raised the unresolved question of who should be responsible when an AI recommendation contributes to an error. The report did not describe a particular incident or establish a legal answer.

Workplace AI Takes Several Forms

Khan’s remarks focused on how technology affects work, rather than only whether it replaces workers. Electronic monitoring can track activity or safety; algorithmic management can analyze interactions and guide workers; and automation can transfer specific tasks from people to software. Her argument was that each can change working conditions, even where a human remains in the job.

The report noted some legislative activity. Khan said laws passed in Illinois and Vermont prohibit replacing mental health workers with AI, while a California law bars companies from advertising an AI chatbot as a health professional. She also said a California bill under consideration would prevent workers from being required to train their replacements, including an AI replacement. The report did not give the bill’s number, status beyond being considered, or further details of the laws.

“Even if they don’t lose a single job to AI, it doesn’t mean there are no harms to workers.”

— Mishal Khan, senior researcher at the UC Berkeley Labor Center

How Widespread Are These Risks?

The summit report describes expert concerns and examples, but does not quantify how often AI is causing de-skilling, surveillance, wage pressure or changes in patient relationships across healthcare. It also does not provide study methods or estimates of how many workers or patients may be affected. The risks Khan outlined should not be read as proof that every AI system produces these outcomes.

Details about accountability for AI-related errors remain unsettled in the report. It does not specify which parties would be legally responsible in particular cases, or how that responsibility might vary by system and setting. The status and scope of the California bill mentioned by Khan also require further verification.

Policy and Workplace Choices Ahead

The next developments to watch include the progress of the California bill and how states apply laws governing AI in mental health services and health-related chatbots. The report does not give a timeline for legislative action or identify upcoming votes.

Healthcare organizations’ choices about monitoring, documentation, task automation and worker input will shape whether the risks described at the summit materialize. Further reporting would need to establish how widely these tools are used, what data they collect, how workers can challenge automated decisions, and what safeguards preserve clinical skills and patient relationships.

Key Questions

What did the expert warn about?

Mishal Khan warned that AI can harm healthcare workers even without eliminating jobs, including through surveillance, de-skilling, reduced autonomy and work speed-up.

Did the report say AI is already replacing healthcare workers broadly?

No. Khan discussed possible harms and legislative responses, but the report did not provide evidence that AI is replacing healthcare workers broadly or quantify job losses.

What example did Tina Hernandez-Boussard give?

She described medical residents at a hospital in Uganda who lost access to electricity and had no AI to read electrocardiograms. She said they lacked confidence synthesizing information to make recommendations.

What does ‘gigification’ mean in this report?

Khan used the term for a possible shift toward algorithmically managed work, such as nursing platforms where workers bid for shifts and algorithms may influence pay. She presented this as a warning about a potential development, not a confirmed widespread practice.

What remains unknown about AI’s effects in healthcare?

The report does not establish how common the listed harms are, how many workers or patients are affected, or who would be liable in specific cases involving an AI recommendation.

Source: rss