SRPD Flock Camera Community Meeting

These are the ten questions we're handing out at the meeting. This page provides the background behind them: where the numbers and claims come from, what SRPD's policies say, and why each question is worth answering publicly.

The questions are not based on an assumption of bad faith. They come largely from SRPD's own published material. They focus on what the department's transparency portal and ALPR policy do, and do not, tell the public.

What the portal reports

As of September 8, 2026, SRPD's Flock transparency portal listed the following figures. Detection, search, and hit counts cover a rolling 30-day period and change daily, so these numbers should be understood as a snapshot from that date.

1. Will SRPD shorten its 30 day retention period, and to what?

Flock announced on August 13, 2026 a seven day default retention period, saying over 90% of searches without a full plate are done within a week. That is narrower than it sounds. It describes searches run without a full plate number, not searches in general. Flock also says existing customers keep their current retention periods.

Seven days is not the figure privacy advocates have been asking for. The ACLU's Automatic License Plate Reader Privacy Model Bill, published in February 2026, says captured plate data “shall be deleted after 48 hours,” and then lists the exceptions: data needed as evidence of specific unlawful conduct, or relevant to an ongoing felony or missing persons investigation. The point of the 48 hours is that everything else goes.

Writing for the ACLU in August 2026, Chad Marlow described Flock's proposal as possibly a step in the right direction, while noting it does not reach that 48 hour mark.

SRPD's transparency portal lists the department's retention period as 30 days. Flock's announcement does not, by itself, change that number.

So the question is not simply whether Flock reduces its default from 30 days to seven. It is how long Sonoma County agencies believe they need to retain records of people who are not suspected of a crime. Every additional day keeps more innocent people's movements available for future searches. Santa Rosa currently keeps that data for 30 days. Why 30 days? And what evidence supports keeping it that long?

2. When Evidence Mode is used, exactly what is preserved, for how long, and who decides what qualifies?

Flock says data associated with a case can be preserved rather than aging out after seven days.

Writing for the ACLU in August 2026, Chad Marlow raised a question about what triggers that preservation. Does Evidence Mode hold only information investigators have determined may be evidence in a specific case, or does the act of searching cause information to be preserved?

The distinction decides what the seven day default is worth. If only actual evidence is kept, the change could meaningfully reduce unnecessary retention. If a search alone can preserve broad amounts of ALPR data, the default protects considerably less than it appears to.

3. Of the 755,377 unique vehicle detections reported in the last 30 days, how many were connected to an actual criminal investigation?

Those detections are not 755,377 suspects. They are vehicles the system recorded, which in a city of this size means a large share of everyone who drove anywhere that month.

Flock may well have helped solve crimes here. What SRPD can describe, and the public cannot, is how much information about people suspected of nothing had to be collected to produce those results.

4. Of the 44,020 hotlist hits reported in the last 30 days, how many were confirmed matches?

An ALPR alert is not proof that a vehicle is wanted. SRPD's own policy recognizes this, requiring officers to verify an ALPR response before taking enforcement action on the strength of an alert alone. A raw count of hits therefore says nothing about how accurate the system is.

The portal names two hotlist sources behind those alerts: California SVS and NCMEC Amber Alert. Roughly 44,000 hits a month against two lists is worth an explanation on its own.

A transparency report that answered the question would show:

• total alerts;
• confirmed matches;
• false positives;
• outdated or invalid records; and
• enforcement actions resulting from confirmed matches.

5. How many outside law enforcement agencies can access or search what Santa Rosa's cameras collect?

SRPD's transparency portal lists a large network of law enforcement organizations granted access to Santa Rosa data.

That changes the shape of the question. This is not a set of cameras operated by the Santa Rosa Police Department for the Santa Rosa Police Department. It is one part of a much broader network, and a resident cannot weigh what that means without knowing who is in it and on what terms.

6. When an officer from another agency searches Santa Rosa's data, whose rules apply: SRPD's, their own agency's, or both?

SRPD's written ALPR policy sets specific requirements for authorized use, training, access, auditing and sharing. It says department members must complete department approved training before accessing ALPR data, and that the data may be accessed only for legitimate law enforcement purposes.

An officer employed by another agency is not an SRPD employee and has not necessarily had that training. Which set of rules governs their use of information that Santa Rosa's cameras collected?

7. If another agency uses Santa Rosa's data in its own investigation, can it keep that information after SRPD's retention period ends?

SRPD's policy says ALPR data should normally be purged once the applicable retention period expires, unless it has become, or is reasonably expected to become, evidence, or is subject to some other lawful requirement.

That makes it matter a great deal who is doing the investigating. Suppose a vehicle is recorded in Santa Rosa on January 1 and SRPD's retention period runs out. An agency somewhere else searches the Flock network on January 20 and uses that record in a case of its own.

Does the outside investigation keep the Santa Rosa data alive? If so, for how long, and under whose retention policy? A resident recorded on January 1 has no way to find out.

8. What prevents an agency with access from using Santa Rosa's data for a purpose SRPD itself prohibits?

SRPD's transparency portal lists prohibited uses, among them immigration enforcement, harassment or intimidation, use based solely on protected class, and personal use.

The network those cameras sit in contains agencies operating under different policies and in different legal environments. Flock has said local agencies will be given more ability to restrict what other Flock customers can search their data for. The ACLU has questioned whether such technical restrictions reliably prevent improper searches, pointing out that when officers were already required to give a reason for a search, some typed "investigation" or even "hehehe".

SRPD should be able to name the mechanism, technical or administrative, that enforces Santa Rosa's restrictions against an agency Santa Rosa does not employ.

9. How many improper Flock searches has SRPD identified, investigated, or disciplined?

SRPD says system access is logged, and its policy calls for regular audits. Flock is also making its Audit Assistance system available across its customer base, which it says flags abnormal search behavior. The ACLU has pointed out that the tool has never been independently evaluated.

That a log exists is worth something. What the audits have found is the part the public can judge, and the only part that shows the oversight works.

10. How many searches produced a useful investigative lead, set against the volume of data collected?

The portal reports 147 cameras, 755,377 unique vehicle detections in 30 days, 817 searches, and 44,020 hotlist hits. It also gives examples of investigations Flock has helped. Those examples are relevant and worth reading.

They do not, on their own, describe how well the system works. Evaluating a surveillance system honestly needs both halves of the equation: what it collected, and what the collection accomplished. Only one half is currently published.