EmpresaEmpleoToluna Start Login
Contacto

Crafting a regulatory agenda for GenAI: physicians’ POV 

Healthcare professionals sit down to discuss the regulation for GenAI's usage in the healthcare industry

Glenna Crooks, PhD and Paul Hambly, EVP of Supply at Toluna

AI is transforming industries across the board, and healthcare is no exception. This is the seventh in a series of articles reviewing physicians’ perspectives on GenAI. This article explores their concerns about regulatory challenges, uncovering an overall lack of confidence in regulatory processes and several features of GenAI technology they believe will challenge regulators.

Regulatory readiness gap

Notably, fewer than half (43%) of physicians are confident that regulatory processes and oversight will succeed in ensuring GenAI is safe to use for patient care. Pediatricians express the least confidence in regulatory processes; only 29% are confident, compared to those who care for adults (44%) or both children and adults (46%).

Rocky regulatory terrain

Driving physicians’ lack of confidence is a concern about GenAI features that they view as troublesome:

  • Pace: 80% are concerned that regulators will not be able to keep up with the pace of GenAI development. Pediatricians are significantly more likely (86%) to hold this view than physicians caring for adults only (79%) or both children and adults. Low-volume physicians (82%) and mid-volume physicians (81%) are more likely to hold this view than high-volume physicians (75%).
  • Unpredictability: 65% are concerned that GenAI results are inconsistent: the same prompt can result in different responses; GenAI outputs cannot be predicted like those of other software.
  • Accountability: 80% cite GenAI’s «black box» nature makes it difficult to assign responsibility if a patient is harmed. Pediatricians (87%) are more likely to be concerned about this view than physicians caring for adults only (79%) or for both children and adults (80%). Low-volume physicians (83%) and mid-volume physicians (80%) are more likely to hold this view compared to high-volume physicians (75%).
  • Privacy: 75% worry about patient-identifiable data entering training datasets. Pediatricians (83%) are more likely to hold this view than physicians caring for adults only (74%) or for both children and adults (73%).
  • Bias: 64% fear that existing healthcare biases will be amplified by GenAI applications.  Pediatricians (71%) are more likely to hold this view than physicians caring for adults only (65%) or for both children and adults (63%).
  • Security: 81% fear malicious actors will outpace their own use of GenAI, placing medical equipment, life-support systems, and patient privacy at risk. Pediatricians (86%) are more likely to express this concern than physicians caring for adults only (80%) or for both children and adults (81%).
Two healthcare professionals discuss the trust and risk of GenAI adoption in the healthcare industry

Unearned trust risks

One overarching concern stands out for most physicians that is worth special note. 75% agree that GenAI is “built to be fast and convincing, not cautious and accurate.”  Their concern, therefore, is that its «warmth, patience, and kindness lead people to trust it more than they should.» Speed and empathy are features that may challenge regulators and clinicians the most as GenAI applications become part of the healthcare landscape.

Coming next:

Our next article will explore a method that physicians have traditionally used to build trust with patients and peers: the oath. We’ll investigate whether physicians would endorse an Oath for GenAI in healthcare

Endnote:

As part of our “AI Everywhere” strategy, Toluna is committed to helping organizations navigate the opportunities and responsibilities that AI brings with it. We partnered with Glenna Crooks, PhD, a recognized policy strategist in global healthcare, to engage over 2,000 physicians on their views of generative AI.[1]

Using Curizon, Toluna’s proprietary panel of healthcare professionals, we explored perceived benefits and risks of AI in healthcare, accountability in the event of harm, and the need for ethical guidelines. The research also examined the advisability of a GenAI Oath modeled after traditional oaths taken by healthcare professionals.

[1]This survey was scripted and programmed by Toluna and fielded in February 2026 with 2,739 healthcare professionals in Toluna’s proprietary healthcare panel Curizon. Survey author: Perso

También te puede interesar