#fraud-detection

#fraud-detection

The IRS stated that its fragmented landscape can lead to undesirable outcomes including duplication of effort and cost, poor understanding of gaps in coverage, and suboptimal case selection.

Calls where no one responds are rarely accidental. In many cases, they are automated reconnaissance events. Fraud operations run at industrial scale, and before they invest human effort in a target, they validate that a number is active and answered by a real person.

Driving the news: Musk took to X on Friday to celebrate the release of Medicaid data on claims, medical procedures and payments from January 2018 through December 2024. "Medicaid data has been open sourced, so the level of fraud is easy to identify," Musk wrote. "DOGE is not a department, it's a state of mind." The Department of Health and Human Services collects Medicaid claims data from states that administer the program. Until now, much of the information has been fragmented and difficult to access.

At the beginning of the year, I looked more closely at one particular statement than I had before. I was shocked by the number of transactions I didn't recognize. They turned out to be subscriptions. My 17-year-old daughter told me that she'd been offered a special deal at the Verizon store: access to Apple Music for up to six people for $10 a month. She was desperate to take advantage of the promotion and said the streaming service had an amazing selection of songs.

A new investigative report from Reuters, citing Meta's own internal documents, found that the company's platforms show, on average, an estimated 15 billion "higher risk" scam advertisements to its users every day. Reuters reported that a 2024 document showed that Meta makes about "$7 billion in annualized revenue" from those scam ads each year. The internal documents revealed by Reuters show that Meta expected as much as 10 percent of its 2024 ad revenue "would come from ads for scams and banned goods."

Businesses around the world have seen a surge in fake expense receipts created using AI-powered image generators from OpenAI and Google, among others, according to the Financial Times.

Scala is a powerful language that combines object-oriented and functional programming, making it ideal for building high-performance, data-intensive, and enterprise-grade applications on the JVM. With the release of the FraudLabs Pro Scala SDK, developers can now easily integrate fraud detection into their Scala projects using the FraudLabs Pro API. The SDK simplifies implementation, helping businesses protect their applications against fraud with minimal coding effort.

Ad fraud is increasingly infiltrating B2B marketing through fake leads, primarily from bot-generated form fills and fabricated prospect lists, leading to inefficiencies and damaged reputations.

California colleges are implementing AI to combat the rising issue of ghost students—fake or stolen identities used to fraudulently enroll for financial aid.

Crystal Intelligence has received a strategic investment from Tether to enhance its blockchain analytics capabilities, focusing on risk monitoring, fraud detection, and regulatory intelligence.

The biggest misconception about accountants is that accounting is designed to detect fraud. An audit is not designed for that purpose, despite the capability to do so.

Fannie Mae and Palantir Technologies have announced a partnership aimed at utilizing AI to detect mortgage fraud swiftly, reducing detection time from months to seconds.

Card declines, often seen as a simple risk issue, are deeply linked to data mismanagement and can lead to a significant loss of good transactions.