AI Radiology Billing Faces Payer Pushback: What Imaging Practices Should Know

Why Are Payers Scrutinizing AI in Radiology Billing?

As of September 2026, the Blue Cross Blue Shield Association published an analysis estimating that AI-enabled coding tools added $942 million in costs to BCBS plans over 2024 and 2025 without a corresponding increase in the complexity of care delivered. Radiology, where AI adoption for image interpretation and reporting has outpaced every other specialty, is at the center of this scrutiny. Practices using AI to assist with coding, documentation, or report generation need to understand what payers are now watching for.

• The finding: BCBSA estimates $942 million in additional costs from AI-assisted coding, with $653 million tied to secondary diagnosis codes shifting claims into higher-paying categories.
• Why radiology is exposed: More than 63 percent of healthcare organizations now use AI in revenue cycle management, and radiology leads AI adoption for clinical and billing applications.
• What to do: Audit how AI tools interact with your coding and reporting workflow to confirm that every code billed is supported by clinical documentation and physician review.

What the BCBSA Analysis Found

On September 24, 2026, the Blue Cross Blue Shield Association released a claims analysis covering its member plans, which insure more than 100 million Americans. The analysis found that hospitals using AI-enabled documentation and coding tools showed a measurable increase in billing complexity without a corresponding change in the care delivered. BCBSA estimated the additional cost at $942 million over 2024 and 2025, compared to 2023 baseline levels.

Approximately $653 million of that total came from secondary diagnosis codes that moved claims into higher-paying DRG categories. As TechCrunch reported, BCBSA characterized the pattern as AI tools identifying billable conditions that were previously undocumented, rather than patients actually becoming sicker. The association reported that more than 63 percent of healthcare organizations now use AI in revenue cycle management, a figure that roughly doubled from 30 percent in 2023.

For radiology, this finding carries outsized significance. Radiology was among the first medical specialties to adopt AI at scale, initially for image interpretation (computer-aided detection, automated measurements, incidental finding alerts) and increasingly for documentation, coding, and report generation. When payers start auditing AI-assisted billing patterns, radiology claims are among the first they will examine.

How Does AI Scrutiny Affect Radiology Practices Specifically?

Radiology practices face a unique exposure to AI-related payer scrutiny because AI tools in radiology operate at multiple points in the billing chain. An AI system might flag an incidental finding during image interpretation, which triggers a secondary diagnosis code, which shifts the claim into a higher-paying category. If the radiologist did not independently confirm the finding and document it in the report, the code is vulnerable to audit.

The professional component of a radiology claim is based on the physician’s interpretation. When AI assists that interpretation, payers want to see that the physician reviewed the AI output and exercised independent judgment. A report that reads as if it was generated entirely by an AI system, without evidence of physician oversight, is a weaker claim in an audit than one that clearly reflects the radiologist’s own assessment with AI used as a decision support tool.

One question we hear constantly from radiology practice managers is whether AI-assisted reporting changes how the professional component should be billed. As of September 2026, CMS has not issued specific guidance on AI-assisted radiology interpretation and billing. The existing rules apply: the professional component requires a physician’s interpretation and report. How the physician arrives at that interpretation, whether with or without AI assistance, does not change the billing code. What changes is the documentation burden. Payers reviewing AI-flagged claims will look at the report to determine whether the physician actually interpreted the images or simply signed off on an AI-generated output.

AI in Radiology: Which Tools Carry Risk

Not all AI tools in radiology carry the same billing risk. Understanding which tools affect coding and which do not is essential for compliance.

The critical distinction is whether the AI tool influences what codes are billed. Tools that help a radiologist read faster (triage, measurement) do not change the billing profile. Tools that flag additional diagnoses or suggest codes directly change the billing profile and create the pattern payers are now watching for: increased coding complexity without a change in patient acuity.

How to Prepare Your Radiology Practice for AI Audits

Whether your practice uses AI for image interpretation, reporting, or coding, these steps protect your revenue in an environment of increasing payer scrutiny.

1. Map every AI tool to your billing workflow. Identify which AI systems touch any part of the coding or documentation chain. If an AI tool suggests diagnosis codes, flags incidental findings, or drafts reports, document how physician review is applied before codes are submitted.
2. Ensure radiologist sign-off reflects independent interpretation. The radiology report should clearly demonstrate the physician’s own assessment. A report that reads as a template populated by AI software is a compliance risk. The ACR guidelines on structured reporting provide a framework for documentation that supports both quality and billing defensibility.
3. Track your coding complexity trends. Compare your practice’s average CPT distribution and secondary diagnosis frequency against the prior 12 months. A significant increase in higher-paying codes without a corresponding shift in exam mix is the exact pattern BCBSA flagged.
4. Review secondary diagnosis code usage on imaging claims. Every secondary diagnosis code on a radiology claim should trace to a finding documented in the radiology report. An AI-flagged finding that appears in the billing but not in the physician’s report is an audit trigger.
5. Build an audit response process for AI-assisted claims. Have documentation ready that explains your AI workflow, the physician review layer, and how codes are assigned. A clear process documented before an audit is stronger than an explanation assembled after one.
6. Communicate with your billing team about AI-generated codes. Your billing staff needs to know which codes originated from AI suggestions and which came from the radiologist’s interpretation. This distinction matters when a payer challenges a specific claim.

Common Mistakes With AI Billing in Radiology

Across the billing companies we vet, a recurring pattern in radiology is that practices adopt AI tools for clinical efficiency without considering the billing implications. The AI improves read times, catches findings the radiologist might have missed, and generates cleaner reports. All of that is valuable. The problem arises when the AI’s output flows into the billing system without the same level of oversight the practice would apply to a manually coded claim.

The most common mistake is accepting AI-suggested secondary diagnosis codes without verifying that the radiologist documented the finding in the report. An AI system might detect and flag an incidental thyroid nodule on a chest CT, but if the radiologist did not include that finding in the final report, the diagnosis code attached to the claim has no documentation support. When a payer audits that claim, the code is unsupported and the payment is recouped.

The second mistake is conflating AI-assisted interpretation with AI-driven coding. A radiologist who uses AI to help identify a finding and then independently confirms and documents it is practicing evidence-based medicine. A practice that lets an AI system add diagnosis codes to claims without physician review is exposing itself to the exact pattern BCBSA identified. In our experience matching providers with billing partners, the practices that manage AI billing risk best are the ones that treat AI output as a draft, never as a final product.

In-House vs. Specialized Billing Support

The BCBSA analysis adds a new dimension to the question of whether to handle radiology billing in-house or outsource it. Radiology billing already carries unique complexity: professional and technical component splits, modifier 26 and TC usage, contrast and non-contrast code selection, and payer-specific prior authorization requirements for advanced imaging. Layering AI compliance on top of that increases the expertise required to bill correctly and defend claims under audit.

Providers often come to us after an audit has already identified issues with their coding patterns. For radiology practices using AI tools, the audit risk is now higher than it was 12 months ago. A billing partner that understands radiology-specific coding, AI-assisted workflows, and payer audit patterns can catch the issues before they become recoupment demands. Whether your practice handles billing internally or works with a partner, the key question is whether the team managing your claims understands the intersection of AI tools and billing compliance well enough to defend every code they submit.

Frequently Asked Questions

Next Steps

Review how your practice uses AI in the coding and documentation workflow. Confirm that every AI-generated code or finding has been independently reviewed and documented by a physician before submission. For billing support that understands radiology-specific coding, modifier 26/TC splits, and AI compliance, we match imaging practices with billing companies that specialize in radiology across all 50 states.