GAO Warns AI Efficiency May Clash With Safety and Fraud Risks

Why it Matters

Artificial intelligence (AI) is rapidly transforming how doctors document patient care and how hospitals bill for services. The Government Accountability Office (GAO) released a report July 16 examining AI-powered tools that automatically generate medical notes and assign billing codes, a technology gaining traction across U.S. health systems. These tools promise to ease clinician burnout by automating tedious paperwork, but they also introduce serious risks: AI systems can fabricate medical information, introduce bias across patient populations, and potentially enable billing fraud. As hospitals and health systems deploy these technologies with minimal federal oversight, policymakers face urgent questions about how to balance innovation with patient safety and billing integrity.

The Big Picture

Medical documentation consumes enormous amounts of clinician time. Doctors spend time drafting visit summaries, discharge notes, and progress notes. Then comes the billing side: medical coders must assign standardized diagnostic and billing codes to each patient encounter for reimbursement. Both tasks are time-consuming and contribute measurably to clinician burnout, a crisis affecting physician mental health and patient care quality across the country.

AI medical coding and documentation tools promise relief. These systems use large language models and natural language processing to read clinical text or process real-time speech from patient-clinician conversations. Some tools transcribe and summarize encounters using speech recognition combined with AI language models. The technology can reduce documentation burden, enable faster and more consistent coding, and potentially improve billing accuracy. A growing number of health systems and hospitals have already begun deploying ambient AI scribing tools, and several major electronic health record vendors and third-party companies now offer AI medical documentation and coding products.

For clinicians drowning in administrative work, the appeal is straightforward. For hospital administrators, the efficiency gains translate to cost savings and faster billing cycles. The problem is what happens when these systems fail.

AI medical documentation tools can generate inaccurate or even fabricated clinical information in medical notes, creating direct patient safety risks. A clinician who relies on an AI-generated note without adequate review might miss a critical detail or act on false information embedded in the record. That risk is compounded by bias; AI models trained on non-representative data may perform differently across patient populations, potentially delivering lower-quality documentation or coding for certain groups of patients.

Another potential risk is billing fraud if AI systems were to miscategorize diagnoses or procedures, resulting in overbilling or underbilling. Data privacy adds another layer of concern. Recording patient-clinician conversations for AI medical documentation tools requires informed consent and careful handling of protected health information. Patients may not fully understand what happens to recordings of their private medical conversations, or they may not have meaningful choice in whether their conversations are recorded at all.

The Bottom Line

The GAO's analysis arrives as AI adoption in healthcare accelerates. Health systems are not waiting for federal guidance before deploying these tools. Every hospital that implements an AI medical coding system or ambient AI scribing tool without clear federal standards is, in effect, conducting an uncontrolled experiment on patient safety and billing integrity. The report fails to provide specific recommendations; that responsibility falls on Congress and the agencies themselves.

Congress now faces a choice. Policymakers can establish clear standards for validation, bias testing, transparency, and liability before AI medical documentation becomes ubiquitous, or they can allow fragmented state-by-state regulation and individual health system policies to determine how these powerful technologies are deployed. The GAO report makes clear that the current path of minimal oversight introduces unacceptable risks, but the question is whether Congress will act on that warning before the technology becomes too embedded to regulate effectively.

Access the Legis1 platform for comprehensive political news, data, and insights.

Spot something wrong? Report an issue with this article