Key Takeaways
- AI‑driven ambient documentation tools are widely adopted in U.S. and European hospitals, yet many patients are not informed that the technology is being used during visits.
- Proponents say the tools save clinicians time and can refocus physician attention on patients, but concerns remain about accuracy, oversight, and privacy.
- Dr. Kirsten Ostherr of Rice University’s Medical Humanities Research Institute warns that AI models can “hallucinate” incorrect data and that tech companies lack the ethical oath that guides physicians.
- Houston Methodist and MD Anderson Cancer Center confirmed use of ambient AI tools, emphasizing physician review and patient‑safety safeguards.
- Ostherr calls for greater patient and clinician involvement in the design and governance of AI note‑taking systems to ensure they become truly patient‑centric.
Introduction to the Issue
As artificial intelligence continues to permeate everyday life, its role in healthcare—particularly in creating and maintaining medical records—has drawn both optimism and apprehension. Dr. Kirsten Ostherr, director of Rice University’s Medical Humanities Research Institute, told ABC13 that AI note‑taking is “incredibly prevalent” across most hospitals and healthcare systems in the United States and Europe. She noted that while the technology promises to streamline documentation, its deployment often occurs without clear patient consent or awareness.
How the Technology Works in Clinical Settings
Ostherr explained that the AI tools typically operate through smartphone apps or ambient listening devices that capture clinician‑patient conversations in real time. “Usually, physicians are expected to let you know that it’s being used,” she said, “but in practice, that often is not the case.” The software transcribes spoken words into structured notes, aiming to relieve doctors of the burden of manual charting.
Reported Benefits for Clinicians and Patients
One of the primary advantages cited by users is time savings. Ostherr highlighted that doctors have reported the technology “saves them time,” allowing them to allocate more effort toward direct patient interaction. She added, “The promise of these tools is that they really will get doctors’ attention back on the patient.” Some patients who are familiar with the technology have reported feeling “more seen and heard” because their provider is not constantly looking at a screen or notepad during the appointment.
Concerns About Accuracy and Oversight
Despite the potential benefits, Ostherr cautioned that the technology currently lacks sufficient oversight. She stressed the need for validation: “What we really want to see is validation that errors are not being introduced when time is saved.” Without rigorous checks, there is a risk that the convenience of AI could compromise the fidelity of medical records.
The Problem of AI “Hallucinations”
A specific technical risk Ostherr highlighted is the tendency of AI models to produce “hallucinations”—fabricated or incorrect information that appears plausible. “AI note‑taking models can exhibit technological ‘hallucinations,’ leading to incorrect information being documented in patients’ records,” she warned. Such errors could have serious clinical implications if they influence diagnosis, treatment plans, or billing.
Privacy and Data Governance Challenges
Beyond accuracy, Ostherr raised concerns about patient privacy and the broader data ecosystem. She observed, “There is a real concern that this is part of an overreach of recording devices that are in social life that are producing data sets that are out there, and we don’t know who has access to them or what they’re using them for.” The ambient nature of the tools means they continuously capture audio, raising questions about data storage, sharing, and potential secondary uses beyond the clinical encounter.
Ethical Divergence Between Medicine and Tech
Ostherr contrasted the ethical foundations of medicine with those of the technology industry. “Physicians abide by an oath to take care of patients,” she said, noting that technology companies do not have an equivalent oath guiding their product development and deployment. This misalignment can lead to situations where commercial interests outweigh patient safety or privacy considerations.
Hope for a Patient‑Centric Future
Despite her criticisms, Ostherr remains optimistic that AI can evolve into a tool that genuinely serves patients—if stakeholders change their approach. “I think there’s an opportunity for this to become a really patient‑centric and patient‑serving tool,” she remarked. She believes that realizing this potential will require patients and clinicians to play a greater role in the design, testing, and governance of AI note‑taking systems.
Responses from Houston‑Area Hospitals
To gauge local adoption, ABC13 reached out to several Houston‑area hospitals; only Houston Methodist and MD Anderson Cancer Center responded by publication time. Houston Methodist confirmed the use of “ambient AI documentation tools” to support clinicians and enhance the patient experience. A statement from the system emphasized that “every note is carefully reviewed, edited and approved by the physician, who they say remains responsible for all documentation and patient care choices,” adding that “patient safety and quality are our highest priorities, and we have strong safeguards and oversight in place to ensure these tools are used responsibly and appropriately.”
MD Anderson Cancer Center likewise affirmed its use of “ambient listening” tools, explaining that the technology “allow[s] clinicians to focus their conversations with patients, rather than try to write everything they say down.” An official at the center noted that the technology is common and claimed that at least three hospitals in the Houston area employ similar systems.
Conclusion and Call to Action
The conversation around AI‑driven medical note‑taking reflects a broader tension between innovation and responsibility in healthcare. While the technology offers tangible efficiencies and the potential to restore clinician‑patient engagement, it also introduces risks related to accuracy, privacy, and ethical oversight. As Dr. Ostherr’s research underscores, realizing the promise of AI in medical documentation will demand transparent communication with patients, rigorous validation of AI outputs, robust regulatory frameworks, and meaningful involvement of both clinicians and patients in shaping how these tools are built and deployed. Only then can AI become a trustworthy ally in the pursuit of better, safer, and more compassionate care.
https://abc13.com/post/expert-weighs-artificial-intelligence-commonly-used-create-maintain-medical-records/19709559/