Audit raises questions about AI in Medicare billing

Audit raises questions about AI in Medicare billing

22 Sep 2026

An Australian National Audit Office (ANAO) review has raised concerns about the use of AI-enabled software in Medicare billing and compliance.

The report, Artificial Intelligence and Medicare Benefits Integrity, said Services Australia had raised concerns about healthcare providers using AI to check eligibility for multiple Medicare Benefits Schedule (MBS) items and suggest billing options.

These concerns were passed to the Department of Health, Disability and Ageing (DoHDA). The ANAO found the department had not yet fully identified and managed the risks linked to AI use.

An internal report prepared by the department’s Benefits Integrity Division in May identified prompted billing software as an emerging issue. It also identified four commercial tools being used to analyse consultation notes.

Officials were unable to determine which practices or healthcare professionals were using these AI functions. However, the report indicated that most inappropriate claims should be identifiable through existing compliance systems.

The ANAO noted that proposed work to better identify the use of these tools had not been completed at the time of its assessment.

Dr Tony Bayliss, Chair of the RACGP’s Regulation and Compliance Working Group, said the findings highlighted the need for human oversight when using AI for billing.

He said GPs remain responsible for the MBS item numbers included in their claims and that this responsibility cannot simply be handed over to AI.

Dr Bayliss also warned that AI-generated errors could create medico-legal risks if billing advice was accepted without proper review.

The ANAO found that DoHDA’s management of AI in Medicare integrity was partly effective but said a more proactive approach was needed.

The use of AI was described as limited. However, between July 2024 and December 2025, the department used an AI-enabled model that was involved, at least partly, in identifying eight potential fraud or non-compliance cases involving about $5.2 million.

Dr Bayliss said greater engagement with the medical profession would be important when introducing AI into Medicare compliance.

He also said incorrect Medicare claims often result from the complexity of the system and MBS item numbers, rather than intentional non-compliance.

The RACGP has also raised concerns that AI could incorrectly identify appropriate billing as irregular and create additional compliance work for GPs.

The ANAO made three recommendations to DoHDA: assess AI risks with other relevant organisations, strengthen compliance processes, and improve assurance measures, including cybersecurity policies covering AI-specific risks.

DoHDA agreed with all three recommendations. It said AI presents both opportunities and emerging risks across healthcare and government and that it was continuing to strengthen its governance, risk management and oversight arrangements.

The RACGP had previously called for caution around the use of AI for Medicare compliance. In a submission to the ANAO, the college said compliance systems should recognise differences between practice arrangements and warned against relying only on government-set standards. 

Audit raises questions about AI in Medicare billing

Source: Australian National Audit Office (ANAO)