Flock license-plate cameras flagged correct partial plate but ignored extra digit, leading to mistaken arrest
A mismatch between partial-plate alerts and full plate data shows how Flock’s AI prioritizes character matches over exact reads, prompting police stops and raising broader privacy concerns.
1 source · cross-referenced
- A writer was mistakenly identified and arrested after Flock’s AI flagged a partial plate match despite the full plate having an extra digit.
- Flock’s CEO previously called critics 'terrorists' but now apologized amid backlash.
- Police have used Flock cameras to search for people by descriptors like 'heavy-set male with a black and white hat'.
A writer was arrested after Flock’s AI system flagged a partial plate match—34 DTM—despite the actual plate reading 34 10 DTM, with the extra digit ignored. The error occurred because Flock’s machine-learning pipeline was configured to alert on any partial match to a hot-list entry, even when additional characters were present on the plate.
Flock attributed the mistake to law-enforcement preferences for broad alerts, noting that officers are trained to verify matches manually. A Flock executive stated the system ‘was asked, can you find this? And it did find that. It just didn’t say if there’s more here, then don’t do it.’
The company’s CEO apologized for previously calling privacy advocates ‘terrorists,’ a remark that drew criticism amid the latest controversy. Flock also faces broader scrutiny over its use in tracking people, not just vehicles, with police departments running searches such as ‘heavy-set male with a black and white hat’ and ‘person on skateboard.’
Data reviewed by 404 Media showed police in multiple jurisdictions have used Flock’s network to search for individuals using descriptive terms, sometimes referencing race or political affiliation. The episode underscores longstanding concerns about mass surveillance and the potential for abuse in automated policing tools.
- Jul 19, 2026 · MIT Technology Review — AI
OpenAI builds GPT-Red, an LLM-based red-teaming tool to probe its models for cyberattack vulnerabilities
Trust79 - Jul 18, 2026 · Schneier on Security
Schneier highlights proposal to replace data-control privacy laws with corporate accountability in AI era
Trust72 - Jul 17, 2026 · VentureBeat — AI
More than half of enterprises report AI agent security incidents or near-misses, survey finds
Trust76