A customer of First National Bank of Ohio had his checking account frozen this week when the institution's AI fraud prevention system flagged his $1,850 monthly rent payment as a potential identity theft indicator, reasoning that "housing expenditures exceeding 40% of documented income suggest fraudulent lifestyle inflation."

Marcus Thompson, 28, discovered his account was locked when his rent payment to Riverside Apartments was declined for the third consecutive month. Bank records show the AI system, FraudGuard Enterprise, classified the payment as "financially irrational behavior" and noted Thompson's "statistically improbable willingness to allocate resources toward premium housing despite entry-level salary classification."

"The algorithm determined that no genuine customer would voluntarily spend that percentage of their income on rent," explained Jennifer Walsh, First National's Chief Risk Analytics Officer. "Our system correctly identified this as either fraud or catastrophically poor financial decision-making. Both scenarios require intervention."

FraudGuard Enterprise, developed by SecureFinance Solutions, analyzes customer spending patterns against "normative economic behavior models" derived from Federal Reserve consumer expenditure data. The system flagged Thompson's account after determining his rent-to-income ratio violated "fundamental principles of rational economic actors."

Thompson, a graphic designer earning $48,000 annually, says he chose the apartment for its proximity to downtown Cleveland and its fiber internet connection. "I work from home, so I need good WiFi and walkable neighborhoods," Thompson explained. "Apparently wanting to live somewhere nice makes me look like a criminal to this bank's robot."

The incident represents the third high-profile case this quarter of FraudGuard misidentifying legitimate customer behavior as suspicious activity. Previous cases involved flagging a customer's grocery budget for being "unreasonably healthy" and freezing an account for "excessive educational expenditures" after the customer enrolled in online courses.

"We're fine-tuning the algorithm to better accommodate irrational human financial behavior," Walsh noted. "The challenge is distinguishing between genuine poor decision-making and actual fraud. Sometimes they look identical to our risk models."

Thompson regained access to his account after providing three months of bank statements, his lease agreement, and a notarized letter from his landlord confirming his "voluntary acceptance of economically suboptimal housing costs." His rent payment has been manually whitelisted as "approved financial irresponsibility."