AI in Healthcare: Thousands of Doctors Use ChatGPT to Respond to Patients, Raising Safety and Ethical Concerns
Staff Editor
2 min read
Thousands of Doctors Using AI to Communicate with Patients, Raising Concerns
In a significant shift in healthcare communication, thousands of doctors are now utilizing artificial intelligence (AI) to draft messages to patients, sparking debates about safety and ethics in medical practice.
The MyChart platform, widely used for doctor-patient communication, has introduced an AI-powered feature that generates responses to patients’ inquiries. Approximately 15,000 physicians have adopted this technology, which employs OpenAI’s GPT-4 model to simulate the doctor’s communication style.
However, this development has raised alarm among medical professionals and ethicists. A recent case highlighted the potential dangers when AI reassured a patient about a vaccine without access to their medical records. Critics argue that such AI-generated advice could introduce dangerous errors into patient care.
A July study found “hallucinations” or fabricated information in 7 out of 116 AI-generated messages, while a separate investigation revealed repeated errors by GPT-4 in patient communications. These findings underscore the high stakes of AI errors in medical advice.
Despite these concerns, the MyChart tool does not currently disclose when responses are AI-generated. This lack of transparency has led to ethical debates about the role of AI in replacing direct doctor-patient communication.
While proponents argue that AI could save valuable time for overworked healthcare professionals, skeptics warn that the accuracy of medical advice could be compromised. There are fears that doctors might overlook AI-introduced errors, potentially leading to serious consequences for patient health.
The regulatory landscape surrounding AI in healthcare communication remains unclear. Currently, there are no federal rules mandating the labeling of AI-generated messages, leaving patients potentially unaware of the source of their medical advice.
As the debate continues, the medical community grapples with balancing the efficiency AI offers against the paramount importance of accurate, personalized patient care. The ongoing discussion reflects broader concerns about AI’s role in critical sectors like healthcare, where errors can have life-altering consequences.
