Flock says its new tool will help identify police abuse, but hasn’t explained how it works

August 13 2026 – San Francisco – Surveillance‑software firm Flock announced on Tuesday that it is rolling out a mandatory feature called Audit Assistance for every customer in its portfolio. The company claims the tool has already helped “catch abuse” within police departments that use its video‑analytics platform. However, Flock has provided no technical details or independent validation, prompting civil‑rights groups and journalists to demand transparency. The lack of clarity raises questions about accountability, especially as municipalities nationwide adopt AI‑driven policing solutions.

Key takeaways

  • Flock makes Audit Assistance compulsory for all existing and new contracts.
  • The firm says the tool has already identified police misconduct cases.
  • No public documentation explains the algorithmic methods behind the feature.
  • Experts warn the opacity could hinder oversight of law‑enforcement AI.

Background

Flock, founded in 2018, supplies AI‑powered video analytics to over 300 law‑enforcement agencies across the United States. Its products flag suspicious behavior, generate real‑time alerts, and store footage for later review. In recent years, the company has faced scrutiny after several high‑profile incidents where its software was accused of bias. The new Audit Assistance tool arrives amid a broader push for algorithmic transparency, highlighted in recent legislation such as the AI Accountability Act.

The announcement was posted on Flock’s corporate blog and immediately picked up by tech outlets, including a report on TechCrunch. While the company frames the feature as a safeguard for citizens, it stopped short of describing the data sources, model architecture, or audit procedures that underpin the system.

What happened

On August 13, Flock emailed its client base a notice stating that Audit Assistance would become a “standard component” of all deployments starting 30 days later. The notice cited “early successes” in several precincts, where the tool allegedly flagged unlawful stops and excessive force incidents. No case studies, timestamps, or statistical outcomes were shared, and the company declined to answer follow‑up questions from the press.

Simultaneously, advocacy groups such as the American Civil Liberties Union filed a request for information under the Freedom of Information Act, asking Flock to disclose the model’s training data and validation metrics. The firm responded that the details are “proprietary” and that releasing them could compromise security.

Why it matters

Without transparent methodology, auditors cannot verify whether Audit Assistance truly reduces misconduct or merely replicates existing biases. Researchers in the technology sector warn that opaque AI tools can become “black boxes” that shield agencies from scrutiny. As municipalities allocate millions of dollars to AI‑enabled policing, taxpayers deserve assurance that public safety tools are both effective and fair.

The controversy also intersects with broader debates on surveillance. Earlier this year, a story about “'Squashed poo' a sign of water voles' success” illustrated how seemingly innocuous data can reveal hidden ecological patterns, underscoring the power—and responsibility—of data analytics. Likewise, the rollout of Audit Assistance could set a precedent for how law‑enforcement agencies adopt AI, influencing future policy decisions.

What happens next

Flock has scheduled a webinar for September 5 to demonstrate the tool’s user interface, though the agenda lists only “high‑level overview” and “client testimonials.” Industry analysts expect the company to face pressure from city councils that have begun drafting AI‑audit ordinances, similar to those discussed in the recent article “Three clubs, six derbies as Welsh rivals collide.”

Legal experts predict that the lack of disclosed methodology could trigger lawsuits alleging violations of the Fourth Amendment. Meanwhile, civil‑rights watchdogs plan to convene a public forum in October, inviting community members, technologists, and police representatives to discuss the implications of mandatory AI audits.

Frequently asked questions

How is Audit Assistance different from Flock’s existing analytics?

Audit Assistance is marketed as a post‑event review layer that flags potential abuse after an incident, whereas the core analytics provide real‑time alerts during police encounters.

Will the tool be available to the public or only to law‑enforcement agencies?

Currently, the feature is limited to Flock’s law‑enforcement customers; the company has not announced any plans