Flock Safety's AI surveillance tools face new safeguards amid abuse concerns
Flock Safety introduced AI-driven search features that let officers scan camera footage by description, but critics say the safeguards still allow misuse.
Flock Safety’s new suite adds AI-powered watchlists that allow officers to define an area and continuously scan camera feeds for individuals fitting a textual description, moving beyond traditional license-plate searches. The rollout has sparked criticism after incidents of officers using the system to track personal contacts, leading to contract cancellations and a congressional letter to CEO Garrett Langley over a Texas deputy’s 83,000-camera query.
In response, Flock pledged to reduce default data-retention periods, require case codes for each query and implement automated auditing, though the core moderation logic remains on the company’s servers. Experts note that the AI’s content filters block certain categories like race or religion but only warn on political expression, and that the model’s decisions are opaque and untested at scale. Audit logs will record searches and officer actions, but the effectiveness of these logs depends on departmental oversight. The company says the new safeguards will be mandatory by year’s end, yet critics argue they address “bad apple” behavior without curbing the broader surveillance capabilities.
Why it matters
The tools could expand police surveillance while still lacking reliable safeguards against abuse.
In this story
