Why I Was Already Voting No on Flock, Before We Found the Data Breach

Where I started

I want to begin by thanking Chief Richards, the City Attorney's Office, the City Manager's Office, and the staff who worked on this agreement. They took the concerns and redlines I raised seriously, and they secured meaningful improvements.

I also want to be clear about where I started. We know this technology helps solve crimes. We know it has helped take dangerous people off the streets here in Pflugerville. I did not approach this renewal looking for a reason to reject it. I was looking for a contract and a system of safeguards strong enough to justify saying yes.

For me, this renewal had to pass three tests: does the contract adequately protect the City and our residents, has the company earned our trust, and have we responded to the extraordinary level of public engagement and concern surrounding this technology. After reviewing the final agreement and the record before us, I did not believe it passed any of those tests.

Test 1: The contract

The agreement is better than where we started. It requires notice within 48 hours of a suspected cybersecurity incident, annual independent security reports, and a seven-day retention period. Staff deserves credit for securing those improvements. But better is not the same as adequate.

The clearest example is data use. The agreement says Flock may not use our data to train or improve AI models. Then it incorporates Flock's online Privacy Policy, which says captured images may be used as machine-learning training data. If the agreement controls, say so clearly. If the online policy controls, then the protection we negotiated is not real. Pflugerville should not have to guess which promise governs our residents' data.

Flock can also revise that policy simply by updating its webpage. It does not promise to obtain our approval or give us an off-ramp if a material protection is weakened. Flock keeps the power to change the rules whenever it wants, while Pflugerville keeps the obligation to live with them. My cellphone and home internet providers let me cancel my service when they change their terms. Flock won't extend that same protection to a high-risk law-enforcement tool.

Flock's recent PR campaign centered on trust and safety says its customers own their own data, full stop. The agreement we're being asked to sign opens a loophole anyway. The raw images may begin as City data, but Flock defines its own property to include derivative works, analyses, reports, models, and other outputs generated through the service. Owning the input is not the same as controlling it end to end.

And if something goes seriously wrong, the imbalance gets worse. Even where Flock can be held responsible, and only through gross negligence, a high legal bar, its exposure is capped at one year of fees. Flock carries $5 million in insurance. Insurance tells us what coverage may be available. The liability cap tells us what Flock has actually agreed to pay Pflugerville. A serious data breach or cybersecurity incident would cost us substantially more than $5 million, on top of the loss of public trust that would follow.

Flock has capped its risk. Pflugerville taxpayers have not. That is not risk sharing. It is risk transfer, from Flock to Pflugerville residents.

Test 2: Trust

For Flock, trust is a slogan more than something the company has earned. Trust is a judgment based on conduct, and the conduct on the record doesn't support the slogan.

Roseville, California, reviewed 1,427 Flock stolen-vehicle and felony alerts and found plate misreads in 71% (Gizmodo's report on the Roseville records). Officer verification kept those errors from becoming stops or arrests, which credits local safeguards, the kind our own Police Department uses and should keep strengthening. Flock responded by pointing to Roseville's camera configuration and placement. But Flock's own standard terms say it advises on placement, holds final discretion to reject a location, and does not allow customers to reposition its leased equipment. A vendor cannot retain that level of control over placement and then put the accuracy failure entirely on the customer.

Accuracy is only one part of the trust problem. Pflugerville currently allows nearly 80 outside organizations to search our data. At least six of those organizations have documented misuse or abuse on the public record. Kyle PD's own audit logs show outside agencies ran at least 117 immigration-related searches against its Flock data, more than half flagged "ICE" (Texas Observer). A Temple PD officer resigned this August after an enhanced Flock audit flagged a search inconsistent with authorized use. Houston PD is currently investigating officers for policy violations, with some accounts placed on hold during the audit. A former Katy PD officer was charged with stalking and misuse of official information after using Flock to track his ex-wife. Harris County's Sheriff's Office suspended three deputies' Flock access this August after a reporter's inquiry into questionable searches. And the Texas Department of Public Safety, which already has access to our data, ran at least 10 immigration-related searches against Kyle's system flagged as "civil and/or administrative," a category that can involve warrants not signed by a judge, a standard Pflugerville PD does not use for our own searches, which are limited to judicial warrants and public-safety purposes.

I am not claiming that Pflugerville's own data was used in any of those specific incidents. I am saying those incidents demonstrate the risk created when we grant continuous, largely unmonitored access to nearly 80 other organizations without enforceable agreements governing purpose, case justification, auditing, suspension, notice, and consequences.

Kyle offers a nearby warning beyond the immigration-search numbers already cited. Its policy prohibits immigration enforcement outright, and it permanently logs every search. A policy cannot control every outside user. A log can only show what already happened. Neither one stops an improper search before it happens or undoes the harm once it has.

Flock's own expansion strategy matters here too. The company developed and pitched a proposal to turn roughly 350,000 Uber, Lyft, and delivery vehicles into mobile license-plate readers connected to its network, using Nexar dashcams already installed in those vehicles. Flock says it never implemented the proposal. It still developed it and marketed it to law enforcement; at the same time, it was asking cities like ours to trust it with tighter limits and clearer data protections. That tells me something about the company's priorities.

Earlier this year, Flock asked Pflugerville Police for a letter of recommendation and success stories, preferably on official agency letterhead, to support its campaign to return to Austin after Austin ended its contract. The request came after the highly publicized Austin shooting investigation in which Manor's Flock cameras helped locate the suspect. I know this because I requested and reviewed the internal email thread directly. To my knowledge, Pflugerville did not provide the letter. Flock asked an existing public customer to lend official credibility to a vendor campaign aftermath of a tragedy, while the privacy, security, accountability, and contract concerns that led cities to question the platform in the first place remained unresolved. That is not what a company focused on earning back trust looks like, and that gap between what Flock says and what it does is a large part of why I don't believe it has earned Pflugerville's trust.

Test 3: The community

This issue generated more emails to the full Council than any other topic this year. More than 150 people emailed the Mayor and full Council about it. Only about 16 of those emails were neutral or supportive of keeping the system as is.

Community input on this goes far beyond a handful of loud voices online. Between Facebook and Reddit, posts about our Flock deployment have received more than 450,000 views and over 2,000 comments, based on my own review across both platforms. By my estimate, that reflects input from more than 500 people. These are not simply residents who distrust the City or oppose the Police Department. Many support both and don't normally weigh in on City issues at all.

People see the public-safety benefit. They also see a system with enormous reach and too few hard limits. Many residents want us to end the program outright, not renegotiate around the edges. Even a stronger contract would not erase what the community has told us. Flock still holds too much control, and Pflugerville still carries too much risk.

Where this leaves me

Staff did its part. The community did its part. Flock did not do its part.

This technology can help solve crimes. That does not require us to accept an inadequate contract, trust unenforceable promises, or dismiss sustained public concern. On each of my three tests, the contract, the company, and the community, the answer is no.

I will be voting no on the renewal.