The Blue Cross Blue Shield Association (BCBSA) has formally called on the Centers for Medicare & Medicaid Services (CMS) to strengthen protections against Medicare Advantage (MA) fraud. Additionally, it urges the agency to establish clear rules around AI-powered coding tools increasingly used across the healthcare system. These recommendations come as growing evidence shows that bad actors exploit information gaps — and that AI tools may be quietly inflating medical bills.
The Medicare Advantage Fraud Problem
Medicare Advantage serves millions of Americans through private insurers. However, it faces a persistent and growing fraud challenge. Bad actors have found ways to shift billing from original Medicare to MA plans precisely when CMS acts against suspected fraud in fee-for-service settings. Because MA plans receive no real-time notification of payment suspensions, fraudulent providers continue to collect payments. This information gap costs the program billions of dollars annually.
BCBSA’s letter to CMS identifies this gap as one of the most urgent issues needing regulatory action. The association’s recommendations focus on five key areas — each designed to close loopholes that fraudsters currently exploit.
Real-Time Fraud Alerts for MA Plans
Closing the Information Gap
One of BCBSA’s top recommendations is straightforward: CMS should notify Medicare Advantage plans in real time when it suspends payments to a provider over suspected fraud.
Currently, MA plans have no way of knowing when CMS has acted against a provider in fee-for-service. As a result, fraudulent billing simply shifts from traditional Medicare to MA plans. Providers continue collecting payments — just from a different source. Real-time alerts would allow MA plans to act quickly and stop paying problematic providers before losses accumulate.
This reform does not require new technology. It requires CMS to share information it already has with the plans that need it most.
Suspending Payments and Tagging Suspect Claims
Removing Barriers to Fraud Prevention
BCBSA also asks CMS to remove contractual and policy language that forces MA plans to keep paying claims even when fraud is suspected. Under current rules, MA plans may be obligated to pay regardless of whether CMS itself has paid its portion. This creates a situation where insurers must continue funding providers they believe are committing fraud.
The association recommends two practical fixes:
- Tagging suspect claims with a unique code so MA plans can identify them before payment clears
- Pricing suspect claims at zero member liability at the time of a CMS payment suspension
These measures give MA plans the tools to act on suspected fraud in real time rather than attempting costly recoveries after payment.
On Mandatory Payment Suspension Authority
On whether MA plans should receive mandatory suspension authority similar to original Medicare, BCBSA urges caution. Most MA plans currently lack the internal infrastructure for rapid and accurate suspensions. Therefore, a direct application of the original Medicare framework could cause administrative errors. The association suggests a phased or carefully designed approach instead.
Regulating AI Coding Tools in Medicare Advantage
A Growing and Unregulated Risk
AI coding tools have entered healthcare at speed. Hospitals increasingly use AI ambient listening systems to record physician-patient conversations and generate diagnosis codes automatically. A recent BCBSA data analysis attributed $663 million in additional inpatient spending to AI-driven coding inflation — suggesting these tools are raising costs, not just improving efficiency.
BCBSA now asks CMS to impose clear standards on these tools:
- Pre-deployment testing and verification of any AI tool used in MA coding oversight
- Post-deployment monitoring at regular intervals to catch drift or errors over time
- Internal AI governance programs at organizations that deploy these tools
- Standardized disclosure requirements for AI developers, covering tool design, training data, and known limitations
Why Transparency Matters
Without disclosure requirements, MA plans and regulators cannot assess whether an AI tool is designed to optimize revenue at the expense of coding accuracy. The DOJ has already pursued cases where chart-mining AI was used exclusively to find codes that raise risk scores — without any corresponding effort to identify unsupported or overstated diagnoses. BCBSA’s recommendations aim to prevent this pattern from becoming standard practice.
AI in RADV Audits: Support, Not Replacement
Keeping Humans in Charge
Risk Adjustment Data Validation (RADV) audits are how CMS checks whether MA plans have accurately coded the health status of their enrollees. As CMS considers using AI in this process, BCBSA draws a firm line.
The association recommends that AI in RADV audits serve only as decision support — not as a replacement for qualified clinical and coding professionals. Human experts must make final determinations on diagnosis validation and payment recovery.
Furthermore, BCBSA recommends:
- AI-supported RADV workflows should be designed to detect both overpayments and underpayments — not just errors that benefit CMS
- CMS should publish performance metrics for any AI tools used in audits
- Insurers should retain the right to review and appeal AI-assisted findings
These safeguards protect both accuracy and fairness. They prevent AI from becoming a one-directional tool that only recovers money for the government while missing errors that hurt plans or beneficiaries.
A Central Registry for DMEPOS Suppliers
Screening Against Bad Actors
BCBSA also recommends that CMS establish a central registry of non-participating DMEPOS (Durable Medical Equipment, Prosthetics, Orthotics and Supplies) suppliers. MA plans should be able to screen against this registry before processing claims. Currently, plans may unknowingly pay suppliers who are excluded or suspended from original Medicare — another information gap that fraud exploits.
Why These Reforms Matter Now
The Stakes for Medicare Advantage
Medicare Advantage now covers more than half of all Medicare beneficiaries. Fraud and inaccurate coding in this program affect federal spending, insurance premiums, and member trust. At the same time, AI adoption across healthcare is accelerating without matching regulatory oversight.
BCBSA’s recommendations reflect a recognition that the current framework was designed for a different era. Real-time fraud alerts, AI governance standards, and transparent audit processes are not optional improvements — they are necessary updates to protect the program’s integrity.
CMS has signaled awareness of these issues. The agency recently announced plans to expand MA audits and invest in AI tools as part of its anti-fraud strategy. Whether it moves to adopt BCBSA’s specific recommendations remains to be seen.
Conclusion
The Blue Cross Blue Shield Association’s formal request to CMS covers urgent ground: closing fraud gaps, regulating AI coding tools, and ensuring that audits remain fair and transparent. As Medicare Advantage continues to grow, so does the need for a regulatory framework that keeps pace with technology. BCBSA’s five-point agenda offers CMS a practical roadmap — and the pressure to act is building on all sides.
