Tech Policy Press author says AI "synthetic suspicion" may create risks for US policing actions
'Synthetic Suspicion' and the Perils of AI-Prompted Police Actions
Tech Policy Press researcher Jake Laperruque wrote that US law enforcement agencies are using AI-generated flags and scores as the basis for stops, searches and arrests, and that policymakers should address this "synthetic suspicion" before it becomes routine in policing. Cases cited in the article include Border Patrol stopping vehicles based on driving-pattern flags derived from AI analysis of license plate reader data, and ICE using the Palantir-developed ELITE system to select raid locations by confidence score. The author noted that facial recognition mismatches have produced at least three cases in the past year in which the person was held for months; according to Representative Bennie Thompson, ICE treats field biometric matches from the Mobile Fortify app as "conclusive" findings, and officers may disregard evidence of US citizenship such as birth certificates. He argued that automation bias and vendor incentives amplify such risks, and recommended independent review of systems, guidance on the weight of evidence and limits on use cases, but said use should be paused until effective safeguards exist, leaving "synthetic suspicion" with no role in policing.