The Bank Alert Has a Secret Audience

US Border Patrol Predictive Intelligence Targeting Teams, known as PITT units, are reportedly analyzing Americans’ financial activity and sending leads to local police. In one documented case, a Spokane, Washington, team examined Kyle William Olson before officers stopped him while he drove across Montana, ostensibly over an obstructed license plate. Police charged him with driving under the influence and intent to distribute narcotics after finding cannabis products in his vehicle.

A Border Patrol memorandum supplied to Olson said agent Matthew Phelps saw information in law-enforcement-sensitive systems indicating financial patterns commonly associated with illicit narcotics activity. The public reporting does not identify the financial institutions involved, the transactions examined, the database searched, or the rule that converted those records into a lead. That missing mechanism is the central fact, not a footnote.

Banks already generate alerts under financial-compliance rules, while government databases and law-enforcement networks can combine records from multiple sources. It remains unclear whether PITT used bank alerts, purchased data, subpoenas, another government system, or some mixture. The purchase leaves a receipt. The receipt acquires an audience.

Financial-crime detection has lawful purposes. Banks and investigators look for money laundering, fraud, trafficking, and other offenses that often produce recognizable transaction patterns. But a pattern is an investigative claim, not proof of intent. Payees can be mislabeled. Locations can be stale. Shared accounts can join the conduct of several people. A model trained on earlier investigations can repeat their blind spots with cleaner formatting.

The pressure falls on the person stopped. Federal analysts may hold the underlying financial theory, local officers may supply the immediate legal justification, and prosecutors may receive a case whose origin is difficult for a defendant to reconstruct. Surveillance expert Jake Laperruque described the reported practice as parallel construction: concealing the actual source of suspicion behind a different reason for police action. If that account is correct, the system’s central product is durable suspicion with a shortened audit trail.

Review requires more than a general assurance that sensitive systems exist. Courts and inspectors need the query terms, access authority, analyst notes, data provenance, retention periods, model or rule versions, dissemination history, and immutable logs showing who viewed or changed a file. Defendants need disclosure when financial analysis helped cause a stop, along with a route to challenge false records. A warrant requirement may depend on the source and scope of the data, but secrecy cannot decide that legal question for itself.

The immediate questions are answerable if the records survive. Border Patrol should identify the legal authority used by PITT, the number of leads sent to local agencies, how often those leads produced stops or arrests, and whether later evidence validated the financial inference. Without that record, inspectors cannot tell whether the program detects crimes, repeats coincidence, or hides weak suspicion behind an ordinary traffic stop.