Flock's AI surveillance tool faces scrutiny over police misuse and privacy concerns

Flock Safety’s AI-powered surveillance tool, used by police to search for people on camera, faces scrutiny over enabling misconduct and privacy violations. The system allows officers to run searches based on written descriptions, with AI scoring relevance against sensitive categories like race, religion, and political expression. Flock claims safeguards, but experts argue the technology lacks transparency and accuracy, leaving officers to bear error risks. In 2026, Flock announced changes, including shorter data retention and automated auditing, but critics say these measures do not fully address risks. Recent reports reveal at least 50 U.S. officers have been charged with misusing surveillance tools, with 46 cases involving Flock. These include searches for romantic partners, ex-partners, and individuals officers wanted to meet. Tom Bowman of the Center for Democracy and Technology warns Flock’s system may enable policing that violates First Amendment protections. Flock’s moderation system blocks searches for race, religion, and offensive content but issues warnings for political speech, allowing officers to proceed if they acknowledge risks. Kate Ruane of the Center for Democracy and Technology highlights challenges of automated moderation, noting content filtering at scale is inherently flawed. Flock’s AI model, which runs on its servers, scores descriptions against sensitive categories but does not provide officers with confidence scores, leaving them to rely on vague warnings. Despite these limitations, the system remains in use, raising concerns about its role in enabling misconduct and eroding privacy.