Medius Highlights AI-Driven Expense Fraud Concerns with GPT-5.0

Medius has raised concerns about the increasing difficulty finance professionals face in identifying AI-generated fake receipts, particularly with the launch of GPT-5.0. In a press release, Medius shared findings from a survey of 1,000 finance professionals in the US and UK, revealing that 32% of respondents admitted they would not recognize a fake expense report.

The survey also indicated a rise in faked receipts since the introduction of GPT-4.0, with 30% of respondents reporting an increase. This trend is expected to worsen with more advanced AI models like GPT-5.0, which could make fraudulent claims harder to detect.

Gary Hall, Chief Product Officer at Medius, emphasized the need for intelligent anomaly detection systems to combat this issue, as traditional finance systems struggle to cope with AI-generated documents that are indistinguishable from genuine ones. The survey also highlighted inefficiencies in the expense management process, with 45% of respondents citing chasing receipts as a major pain point.