Accounts payable (AP) and accounts receivable (AR) teams are facing a new generation of fraud. Artificial intelligence now enables criminals to create fake vendors, fabricate convincing invoices, and impersonate trusted executives and suppliers through deepfakes—all at a scale and level of sophistication that traditional controls weren’t designed to detect.
According to the Association of International Certified Professional Accountants, the risks extend beyond external attacks. Increasingly, businesses are uncovering fraudulent receipts and invoices created by their own employees.
One common tactic is the use of AI-generated deepfakes to manipulate employees. While many of these fake audio or video messages impersonate company executives, a growing number mimic established vendors. The goal is to convince AR/AP staff to redirect payments to fraudulent accounts.
Criminals are also using AI to package synthetic identities, fabricated invoices and documentation, and fraudulent company websites to create businesses that appear legitimate. As a result, it’s become more difficult for employees to discern a fraud attempt from a genuine business communication.
Evidence of a Trend
These threats represent trends that are impacting consumers and organizations alike.
For example, roughly one in 10 fraud attempts now involves synthetic data. Because these identities aren’t fully tied to a real individual, there is often no legitimate person to detect or report fraudulent activity.
Another growing concern for businesses is the rise of internal fraud. First-party fraud, where customers exploit an organization’s technology or policies for financial gain, has become the world’s leading fraud typology, driving significant losses for merchants.
Lowering Barriers to Entry
As damaging as these attacks can be individually, threats targeting AR/AP functions can have organization-wide consequences by disrupting cash flow and exposing businesses to financial losses.
Perhaps even more concerning is how technology—especially AI—has lowered the barriers to entry for criminals. Even free versions of AI models can generate sophisticated documents, images, and audio with minimal prompting.
This is to say nothing of AI’s ability to sift through vast amounts of data, a functionality that could be exploited by criminals to identify potential targets and uncover organizational vulnerabilities.
AI’s ability to identify security weaknesses has become a top concern for organizations following the release of cybersecurity models like Anthropic’s Mythos, which reportedly uncovered thousands of critical flaws across organizations in multiple industries.
What’s more, the growing AI threat has already prompted companies like Apple to speed up software updates to patch vulnerabilities before cybercriminals can exploit them. According to the tech giant, faster patching has become imperative as AI makes cyberattacks quicker, cheaper, and more scalable.
