User:Louis/Marc Duchen on ALPRs and Flock Safety
 | This is a personal page in Louis's userspace, kept as a reference. It is not a wiki article. |
Marc Duchen represents District 10 on the Austin City Council and has argued for keeping automated license plate readers (ALPRs) in Austin with added safeguards, and for switching vendors rather than ending the program. Each of his arguments is below, in his own words, with the documented record answering it.
Safeguards and vendor oversight can solve it
[edit | edit source]The key difference between my position and advocates that want the city to never use ALPR is that I don't believe this is an insoluble problem. I think there are ways that we can install safeguards, have oversight over our program and vendor, and use the technology in a responsible and helpful way.
Marc Duchen, in a constituent email to a District 10 resident.
The record:
- California State Auditor, February 2020 audited four ALPR agencies and found none had audited their own searches; in Sacramento, 66 percent of searches left the case-number and reason fields blank. information.auditor.ca.gov
- ACLU, July 2, 2026 an Oregon police department was allowed to search Flock after typing "investigation" into the reason field 111 times and "hehehe" 20 times. The safeguard is defeated by nonsense. aclu.org
- 404 Media, December 22, 2025 Flock left livestreams and admin panels for at least 60 of its own AI cameras open on the public internet, where anyone could watch them and pull 30 days of archive. The vendor could not secure its own hardware. 404media.co
- Austin City Auditor, May 2025 Austin's own Flock contract "may leave motorist data open to unintended use," allowing retention past the 7-day deletion rule and sharing with outside agencies outside the Council Resolution. The hole is in Austin's contract right now. austin.widen.net
The question: Which specific safeguard in Austin's program would have stopped the Johnson County, Texas deputy who ran a nationwide Flock search noting "had an abortion, search for female," when Austin's own reason field is the exact kind of gate that reads "hehehe" elsewhere?
"No ALPR ever" is an extreme position
[edit | edit source]In fact, I find the position of "No ALPR ever under any circumstance" a fairly extreme position. That position is essentially giving up and saying we cannot accomplish what most states and major cities have done by balancing public interest and privacy
Marc Duchen, in a constituent email to a District 10 resident.
The record:
- NPR, February 17, 2026 at least 30 localities have deactivated Flock or canceled their contracts since the start of 2025, including the college towns of Flagstaff, Cambridge, Eugene, and Santa Cruz. That is now a mainstream decision, not a fringe one. npr.org
- NPR, February 17, 2026 they pulled out because officials realized their data was being shared far more broadly than they knew, and that federal agencies including U.S. Border Patrol had searched it. They quit after seeing the data flows, not before. npr.org
The question: If ending a city ALPR program is "extreme," why did roughly 30 jurisdictions across the political spectrum reach exactly that decision in a single year once they saw who was querying their residents' data?
Even progressive cities use ALPR
[edit | edit source]including "progressive" cities like New York, San Francisco, San Diego, Seattle, Vancouver, Amsterdam, etc.
Marc Duchen, in a constituent email to a District 10 resident.
The record:
- KQED, June 18, 2026 San Francisco's own police audit found its plate-reader data was improperly accessed hundreds of times by out-of-state and federal agencies since last May, routed through the regional intelligence center, with a third-party group given access "during night hours." One of his model cities only learned this from an audit. kqed.org
- NPR, February 17, 2026 Cambridge, Mass. and Santa Cruz, Calif., both peer progressive cities, are on the list of at least 30 localities that dropped Flock. The "progressive cities use it" list is shrinking. npr.org
The question: If San Francisco is the progressive model, should Austin also expect to discover, only via audit, that its data was queried hundreds of times by federal and out-of-state agencies it never authorized?
165 felonies assisted or solved in the pilot
[edit | edit source]In our one-year pilot APD used the technology to assist with or solve 165 felony (serious) crimes including: Murder, Robbery, Aggravated Assault, Unauthorized Use of a Motor Vehicle, Indecent Exposure, Kidnapping, Evading in a Motor Vehicle, Evading on Foot, Assault with Injury, Stolen License Plates, Possession of Controlled Substance, Violation of a Protective Order, and Sexual Assault.
Marc Duchen, in a constituent email to a District 10 resident.
The record:
- Austin City Auditor, May 2025 the audit says "165 arrests, approximately 133 prosecutions," not 165 solved crimes. The word "solve" never appears in the report, "felonies" is never attached to the 165, and there is no itemized crime-type list tied to it. austin.widen.net
- Austin City Auditor, May 2025 (APD's own response, p. 13) APD itself said the "outcome data lacks significant or complete meaning" because outcomes are "most of which are fully outside of APD's control." The department that ran the tool disowns the outcome numbers. austin.widen.net
- June 3, 2025 Austin City Council work session asked by CM Qadri whether they were "assuming correlation equals causation," APD conceded it is "very difficult to quantify the impact to ALPR being a hundred percent the singular tool." June 3, 2025 work session
The question: The audit says 165 arrests, not 165 felonies solved, and explicitly declines to attribute outcomes to the camera; on what analysis do you upgrade "arrests the tool touched" to "serious crimes the tool solved"?
Without ALPR, victims go without justice
[edit | edit source]The technology is not used for less serious misdemeanors. The absence of this technology now means we lack an important tool to solve serious crimes, and telling victims that we may not be able bring them justice.
Marc Duchen, in a constituent email to a District 10 resident.
The record:
- Austin Police Department, May 13, 2025 Austin cleared 100 percent of its 2023 homicides (75 of 75) and 94 percent in 2024 (67 of 71). The 100 percent year was before the pilot even started. There is no macro room for a camera to have rescued cases that were already cleared. austintexas.gov
- Austin City Auditor, May 2025 with clearance already at the ceiling, the audit could only say ALPR "assisted"; it made no finding that any serious crime went unsolved for lack of it. austin.widen.net
The question: Austin cleared 100 percent of its 2023 homicides with no ALPR; what is the evidence that any specific serious crime went unsolved for lack of it, rather than that being an assertion the audit declined to test?
"The camera found the killer's car"
[edit | edit source]The ALPR searches resulted in the discovery of a vehicle matching that description leaving the area of the shooting within 8 minutes of the murder. The location of the camera was 2 miles away from the location of the murder.
Marc Duchen, in a constituent email to a District 10 resident.
The record:
- City of Austin Office of the City Auditor, May 2025. Austin's own audit of the pilot reports "165 arrests, approximately 133 prosecutions." It never uses the word "solve," never attaches "felonies" to the number, and makes no causation finding. APD's own written response says the outcome data "lacks significant or complete meaning" and the results are "most of which are fully outside of APD's control." [1]
- Austin Police Department, May 13, 2025. Austin cleared 100% of its 2023 homicides (75 of 75) with no ALPR, and 94% in 2024 (67 of 71). A rate already at the ceiling leaves no room for the camera to have rescued cases that were already being cleared. [2]
- Austin City Council ALPR work session, June 3, 2025. Of the two murder cases APD chose to highlight at this hearing, both began with witness or vehicle descriptions, and the 2025 case was arrested by the Kerr County Sheriff using ballistics and a confession, not the camera. June 3, 2025 Austin City Council work session.
The question: This lead began with a doorbell-camera description and closed on forensic evidence matched to the registered owner. What part of that outcome would not have happened through the description-plus-forensics chain that actually solved it?
"ALPR tracked the burglar to Houston"
[edit | edit source]Burglary detectives were able to query the suspect vehicle description and partial license plate through the ALPR system, and they were able to locate the suspect vehicle in Houston.
Marc Duchen, in a constituent email to a District 10 resident.
The record:
- City of Austin Office of the City Auditor, May 2025. The audit reports arrests, not solved-and-attributed cases, and warns the outcome data "lacks significant or complete meaning" and is "outside of APD's control." Locating a vehicle is a lead, not a cleared case, and the audit attributes no single outcome to the camera. [3]
- City of Austin Office of the City Auditor, May 2025. The ~133-prosecution figure is the auditor's own estimate from public-information requests, "a snapshot in time," not adjudicated results tied to ALPR. [4]
The question: "Located the vehicle in Houston" is a lead, not a conviction. Does the audit anywhere establish that this case was solved because of the ALPR query rather than merely with it?
"Manor's cameras stopped the shooting spree"
[edit | edit source]Their 19-hour shooting spree ended when they crossed into Manor, which has deployed ALPRs. They were then caught within an hour.
Marc Duchen, in a constituent email to a District 10 resident.
The record:
- It caught them, it did not prevent them. By his own account the spree ran 19 hours across about 12 locations and injured four people before any camera flagged the vehicle. ALPR is reactive: it helped identify the car after the harm was done and stopped none of the shootings. His own words, track down and arrest, describe an after-the-fact catch, not prevention.
- Austin Police Department, May 13, 2025. Austin was already clearing 94 to 100% of its homicides without ALPR, so a single high-profile multi-jurisdiction event does not show routine citywide surveillance is what drives case outcomes. [5]
- City of Austin Office of the City Auditor, May 2025. Austin's own audit made no causation finding and reports arrests, not outcomes attributable to the camera. [6]
The question: One Manor alert in one weekend is the single strongest case anyone has for ALPR. Is that one event, weighed against millions of daily scans of innocent Austinites feeding a national network, the standard on which the city should re-adopt mass surveillance?
"It's a force multiplier for an understaffed department"
[edit | edit source]It's also important to remember that APD is still short ~350 officers and recently pulled some officers off of specialized work to ensure we have patrols 100% staffed. APD considers tools and technology like this a "force multiplier" that is infinitely more cost-effective than most other tools or personnel.
Marc Duchen, in a constituent email to a District 10 resident.
The record:
- 404 Media (Jason Koebler and Joseph Cox), May 27, 2025. The same networked architecture that lets one officer query thousands of cameras gave ICE "side-door" access: officers nationwide ran more than 4,000 lookups on one Illinois town's cameras, typing "immigration," "ICE," and "ICE WARRANT" into the reason field. [7]
- Electronic Frontier Foundation, October 7, 2025. A Texas (Johnson County) deputy searched 83,345 cameras across 6,809 networks with the note "had an abortion, search for female." A tool that multiplies reach multiplies misuse. [8]
- California State Auditor, February 2020. In Sacramento, 66% of searches left the reason field blank, and none of the four agencies audited its own searches. [9]
- ACLU (Chad Marlow and Jay Stanley), July 2, 2026. An Oregon department was allowed to search Flock after typing "investigation" 111 times and "hehehe" 20 times into the reason field. [10]
The question: If the value is that one officer can reach a nationwide camera network with a single query, how do you contain that identical capability when the querying officer is in another state typing "immigration" into the reason field?
"$114K for 165 felonies is a bargain"
[edit | edit source]I have a hard time understanding anyone arguing the economic tradeoff that our $114K pilot to address 165 serious felonies is not cost-effective. It instead strikes as one of the most cost-effective public safety investments we've ever made
Marc Duchen, in a constituent email to a District 10 resident.
The record:
- City of Austin Office of the City Auditor, May 2025. The numerator of his ratio does not hold up. The audit reports "165 arrests," never "165 felonies solved," attaches no crime-type list, and APD conceded the outcome data "lacks significant or complete meaning." You cannot compute cost-per-serious-felony-solved from a number the audit refuses to call solved. [11]
- City of Austin Office of the City Auditor, May 2025. The $114K figure is not even in the audit; it comes from a separate June 3, 2025 Austin City Council work session. His cost side and benefit side come from two different records, and the benefit side is unattributed. [12]
- Electronic Frontier Foundation, October 7, 2025. The actual objection is not price; it is that Austin's feeds join a national network already searched in a Texas abortion investigation. Cost-effectiveness does not answer a rights problem. [13]
The question: Cost-effective per serious felony solved requires knowing how many the tool actually solved. Since the audit declines to attribute a single outcome to ALPR, what denominator makes your ratio mean anything?
The MOU gate is the gate that already failed in San Francisco
[edit | edit source]Assuming the request is approved, the Lieutenant requires the requestor to sign a Memorandum of Understanding (MOU), which prohibits using the information for any purpose other than the original request and mandates the destruction of the data once it has served its purpose. Additionally, if we find that an outside request is overly broad, we must notify the Office of Police Oversight (OPO), the mayor, and the council.
Marc Duchen, in a constituent email to a District 10 resident.
The record:
- KQED, June 18, 2026: San Francisco ran your exact architecture, a regional intelligence center (NCRIC) as the gatekeeper, and its plate data was still "improperly accessed hundreds of times by out-of-state and federal agencies"; NCRIC gave outside analysts access "during night hours." [14]
- 404 Media, May 27, 2025: the "reason" field that is supposed to police these requests is where officers nationwide simply typed "immigration," "ICE," and "ICE WARRANT"; more than 4,000 lookups gave ICE side-door access with no federal contract. [15]
- California State Auditor, Report 2019-118, February 2020: in an audited program, Sacramento left the reason field blank on 66% of searches, and not one agency reviewed had audited its own searches. [16]
- ACLU, July 2, 2026: an Oregon department was allowed to search Flock after typing "investigation" into the reason field 111 times and "hehehe" 20 times. [17]
- Austin City Auditor (Corrie Stokes), May 2025: Austin's own auditor found the Flock contract "may allow Flock Safety to retain some elements of scan data beyond the seven-day deletion requirement and then share them with outside law enforcement agencies without following Resolution guidelines." [18]
The question: San Francisco had your exact gate, a regional intelligence center and an out-of-state-sharing ban, and its data was still queried improperly hundreds of times "during night hours." What makes Austin's Lieutenant-plus-MOU different in kind, rather than the same gate before its own audit catches up?
"50 to 100% a quarter" is not a number the audit will give you
[edit | edit source]the number of felony arrests and prosecutions increased 50-100% every quarter as adoption increased and police became more familiar with how to use the tool. Terminating the pilot means we actually don't know what the full value or upper limit of the technology really is yet to address serious crimes.
Marc Duchen, in a constituent email to a District 10 resident.
The record:
- Austin City Auditor (Corrie Stokes), May 2025: the audit says only that the figures "increased consistently." It gives no percentage anywhere, covers just three of the pilot's four quarters (March through December 2024), and calls the Flock scan totals unverifiable: "it is impossible to verify the true number of Flock scans." [19]
- Austin City Auditor (Corrie Stokes), May 2025: the same audit could not verify the outcomes at all, and APD's own written response calls the outcome data "lacks significant or complete meaning." [20]
The question: "We don't know the full value" cuts both ways; it means the benefit is unproven. Why resolve that by expanding surveillance instead of first running the causation analysis your own auditor says was never done?
The "purity test" is that a city contract makes your residents queryable nationwide
[edit | edit source]Stopping the city from using ALPR doesn't mean Austin is free from ALPR, it just means city law enforcement can't use this tool. This seems to be me more like a "purity test" than actually stopping surveillance.
Marc Duchen, in a constituent email to a District 10 resident.
The record:
- Virginia Center for Investigative Journalism (WHRO), September 16, 2025: Bridgewater, Virginia, population about 6,600 with five cameras, had its Flock data "accessed ... 6.9 million times over 12 months" by outside agencies across the country. [21]
- The Guardian (reported by The 74), February 10, 2026: one Texas school district's cameras were searched more than 733,000 times in a single month by over 3,100 agencies, with immigration cited as the reason 620 times by 30 agencies, even though the district's own police cannot run immigration searches. [22]
- 404 Media, May 27, 2025: the mechanism is that a local contract feeds a national lookup; local police lookups gave ICE side-door access to a tool it has no contract for. [23]
The question: The difference between a private camera and a city ALPR contract is that the city's feed enters a national lookup queried millions of times by agencies you never authorized. How is declining to enroll your constituents in that a "purity test" rather than the specific harm you are avoiding?
"No human bias" describes the plate read; the danger is what the network does next
[edit | edit source]we use other surveillance technology, including HALO cameras in the downtown area, that unlike ALPR rely on human surveillance - and therefore human bias.
Marc Duchen, in a constituent email to a District 10 resident.
The record:
- Commonwealth v. McCarthy, Massachusetts Supreme Judicial Court, April 16, 2020: the court held that "with enough cameras in enough locations, the historic location data from an ALPR system ... would invade a reasonable expectation of privacy and would constitute a search." The constitutional problem is automated aggregation at scale, not the presence of a human operator. [24]
- 404 Media, May 27, 2025: it is precisely the automated, networked design that let officers in other states reach local cameras and hand ICE side-door access. [25]
- Virginia Center for Investigative Journalism (WHRO), September 16, 2025: that same automation is what let outside agencies query one small town's cameras 6.9 million times in a year, a scale no human watcher could ever reach. [26]
The question: Removing the human removes bias from reading a plate, but the harm is the nationwide retention and sharing the automated system then performs. How does taking the human out make that less dangerous rather than more?
The challenges the pilot revealed are ICE access, an abortion search, and a vendor exposing its own cameras
[edit | edit source]I would flip the entire narrative on its head and instead argue that a pilot is intended to reveal both opportunities and challenges
Marc Duchen, in a constituent email to a District 10 resident.
The record:
- EFF, October 7, 2025: a challenge the network already revealed: a Johnson County, Texas official ran ALPR searches logged "had an abortion, search for female," one of them sweeping 83,345 cameras. [27]
- 404 Media, May 27, 2025: another: local lookups gave ICE side-door access to a nationwide camera network it holds no contract for. [28]
- 404 Media, December 22, 2025: and the vendor itself left at least 60 of its AI cameras open on the public internet, where anyone could watch the livestream, download 30 days of archived video, and change settings. [29]
- 404 Media, June 8, 2026: the trajectory is toward more identification, not less: Leonardo's SignalTrace "would turn ALPR cameras from devices focused on tracking cars to ones that can more readily track the location of particular people." [30]
The question: If a pilot exists to reveal challenges, the ones it revealed are federal side-door access, an abortion-investigation search, and a vendor caught exposing its own cameras. At what point does "balancing" become "declining"?
Popularity doesn't settle a rights problem
[edit | edit source]Many have written to our office in support of ALPRs, and believe we can use technology in a smart and safe way to protect ourselves and solve crimes. Some have employed the same technology in their own neighborhoods.
Marc Duchen, in a constituent email to a District 10 resident.
The record:
- 404 Media (Jason Koebler and Joseph Cox), May 27, 2025: local police across the country ran "immigration" and "ICE" lookups in Flock's reason field, giving federal agents side-door access to a system ICE has no contract for; more than 4,000 such lookups came out of one Illinois department's records alone. [31]
- Electronic Frontier Foundation (Dave Maass and Rindala Alajaji), October 7, 2025: a Johnson County, Texas official ran a plate search logged "had an abortion, search for female," a single query reaching 83,345 cameras across 6,809 networks. Same state as Austin, same Flock network the pilot fed. [32]
- NPR (Jude Joffe-Block), February 17, 2026: at least 30 localities, Flagstaff, Cambridge, Eugene, and Santa Cruz among them, dropped Flock since 2025, many after learning federal agencies including U.S. Border Patrol had searched their data. [33]
The question: When you count your constituents' support, whose interest speaks for the person in another jurisdiction whose movements your cameras feed into a national database queried for immigration and abortion cases?
A statewide statute already failed this exact test
[edit | edit source]The Council passed the Transparent and Responsible Use of Surveillance Technology (TRUST) ordinance in April to provide strict guidance how we adopt surveillance technology like ALPR in the city. City departments must get Council approval for this technology and assess existing surveillance technology.
Marc Duchen, in a constituent email to a District 10 resident.
The record:
- Virginia State Crime Commission, January 2026: Virginia's 2025 law flatly barred sharing ALPR data with federal or out-of-state agencies, a binding statewide statute, stronger than any city ordinance. Within months, 13% of agencies (20 of 159) still gave other states continuous access and 6% (9 of 159) gave federal agencies continuous access; 21% over-retained data and 16% had no written policy at all. [34]
- California State Auditor (Report 2019-118), February 2020: where policies existed on paper they were defeated in practice: the audited agencies "either did not have ALPR policies or their policies were deficient," none audited their own searches, and in Sacramento 66% of searches left the reason field blank. [35]
The question: If Virginia's statewide statute could not stop 20 agencies from handing other states continuous access, what specifically makes Austin's ordinance enforceable where the stronger law was not?
The review only runs one direction
[edit | edit source]I supported the TRUST ordinance and passed an amendment to further ask staff to evaluate, after one year, whether the restrictions in the ordinance inadvertently hinder public safety.
Marc Duchen, in a constituent email to a District 10 resident.
The record:
- California State Auditor (Report 2019-118), February 2020: the documented failure is not that safeguards blocked policing; it is that they were bypassed. In Sacramento 66% of ALPR searches left the reason field blank, and none of the four agencies reviewed audited their own searches. [36]
- ACLU (Chad Marlow and Jay Stanley), July 2, 2026: one Oregon police department was allowed to search Flock in September 2025 after typing "investigation" into the reason field 111 times and "hehehe" 20 times, avoiding the words that would trigger review. [37]
- Virginia State Crime Commission, January 2026: under a binding statewide ban, 13% of agencies (20 of 159) still gave out-of-state agencies continuous access and 6% (9 of 159) gave federal agencies continuous access. The rules were not too strict; they were ignored. [38]
The question: Your amendment asks whether the restrictions impeded public safety; where is the parallel one-year review of whether the program's data was accessed for immigration or reproductive-health investigations, which is the harm the restrictions exist to prevent?
The council signs the Flock contract, not social media
[edit | edit source]ALPR often seems like focusing on a small ice floe when you're already colliding with giant icebergs, in the form of social media, big data, or even just the impact on screens on kids and society.
Marc Duchen, in a constituent email to a District 10 resident.
The record:
- Electronic Frontier Foundation (Dave Maass and Rindala Alajaji), October 7, 2025: the abortion-investigation search is not hypothetical or distant. It happened in Johnson County, Texas, across 83,345 cameras on the same Flock network Austin's pilot fed. [39]
- 404 Media (Jason Koebler and Joseph Cox), May 27, 2025: ICE side-door access came through the same reason-field lookups the city's own contract enables, more than 4,000 in one department's records. [40]
- City of Austin, Office of the City Auditor, May 2025: even the "value" side of the ledger is unproven: the audit made no causation finding, and APD's own response said the outcome data "lacks significant or complete meaning ... most of which are fully outside of APD's control." [41]
The question: The council's remit is the surveillance it authorizes and pays for; how does the existence of bigger private harms you cannot regulate bear on whether the city should run a program already documented enabling ICE and abortion-investigation searches?
"Measurable" is the one word the audit refused to use
[edit | edit source]which despite challenges and risks at least provides a measurable public safety value - something I cannot really say for technology like social media.
Marc Duchen, in a constituent email to a District 10 resident.
The record:
- City of Austin, Office of the City Auditor, May 2025: the audit reported "165 arrests, approximately 133 prosecutions," never used the word "solve," attached no felony crime-type list to the 165, and drew no causation finding. APD's own management response called the outcome data "lacks significant or complete meaning ... most of which are fully outside of APD's control." [42]
- City of Austin, Office of the City Auditor, May 2025 (same report, p. 6 and p. 20): the one large number, the scan count, is an unverified APD estimate: "While it is impossible to verify the true number of Flock scans, we include APD's estimate below for reference." The audit also covered only three of the pilot's four quarters. What got measured was activity, not value. [43]
The question: What is the measurement? The audit counted arrests and scans, not crimes solved because of the camera; which specific figure in the record measures the public-safety value rather than the activity?
A better vendor keeps the same national-sharing surface
[edit | edit source]not throwing the baby out with the bath water would be to find a different vendor that might be more responsible or might address some of the concerns we've heard from the Council and constituents.
June 3, 2025 Austin City Council work session.
The record:
- ACLU, October 24, 2025. The sharing is written into the vendor's standard contract, not a Flock-only bug. Flock's default agreement gives the company the right to share data with federal and local agencies for "investigative purposes" even when a department opts out, under a "worldwide, perpetual, royalty-free" license. ACLU, October 24, 2025
- City of Austin, Office of the City Auditor, May 2025. Austin's own Flock contract had the same language: Section 4.4 grants a "non-exclusive, worldwide, perpetual, royalty-free right and license," which the auditor warned "may allow Flock Safety to retain some elements of scan data beyond the seven-day deletion requirement and then share them with outside law enforcement agencies without following Resolution guidelines." City of Austin, Office of the City Auditor, May 2025
- Commonwealth v. McCarthy, Massachusetts Supreme Judicial Court, April 16, 2020. The constitutional problem is scale, not the logo. The court held that "with enough cameras in enough locations, the historic location data from an ALPR system in Massachusetts would invade a reasonable expectation of privacy and would constitute a search for constitutional purposes." Commonwealth v. McCarthy, Massachusetts Supreme Judicial Court, April 16, 2020
The question: Name the vendor that sells a citywide plate-reader network without the retention-and-national-sharing architecture, because that data model, not the Flock brand, is the objection. Switching the logo keeps the same query surface.
The "hypothetical" risks already happened
[edit | edit source]focus on who is it already helped us bring to justice versus these sort of risks that we've identified and hypotheticals ... what we've already gained or stand to gain by hopefully adding safeguards and continuing this program versus what we may be at risk of losing.
June 3, 2025 Austin City Council work session.
The record:
- Electronic Frontier Foundation, October 7, 2025. Not hypothetical. A Johnson County, Texas sheriff's office ran an ALPR search noting "had an abortion, search for female," a query that reached 83,345 cameras across 6,809 networks in a single month. Same state as Austin, same Flock network the pilot fed. EFF, October 7, 2025
- 404 Media, May 27, 2025. Not hypothetical. Local police gave ICE "side-door access" through Flock lookups; officers nationwide typed "immigration," "ICE," and "ICE WARRANT" into the reason field, more than 4,000 lookups in one department's records. 404 Media, May 27, 2025
- KQED, June 18, 2026. Not hypothetical. San Francisco's plate data was "improperly accessed hundreds of times by out-of-state and federal agencies," routed through the regional intelligence center, some queries running "during night hours." KQED, June 18, 2026
- City of Austin, Office of the City Auditor, May 2025. The "gains" side is the unproven one. The audit reports "165 arrests, approximately 133 prosecutions," never the words "solve" or "felonies," and APD conceded the outcome data "lacks significant or complete meaning." City of Austin, Office of the City Auditor, May 2025
The question: Your ledger puts realized gains against hypothetical risks, but the abortion search, the ICE access, and the San Francisco breach are documented while your gains are the numbers the audit refuses to attribute to the camera. Which column is actually the hypothetical one?
The ordinance regulates surveillance, not streetlights
[edit | edit source]I worry that as it's currently designed, it's going to have some unintended consequences that may stymie our ability to adopt basic technology for everyday quality of life issues that we deal with.
April 23, 2026 Austin City Council meeting.
The record:
- City of Austin, Office of the City Auditor, May 2025. The ordinance's bars target documented misuse, not conveniences. Austin's Council Resolution bars using ALPR data to investigate "First Amendment activities, reproductive care access, and immigration status." City of Austin, Office of the City Auditor, May 2025
- Electronic Frontier Foundation, October 7, 2025. "Reproductive care access" is exactly the category Flock data was used for elsewhere: the Johnson County abortion search across 83,345 cameras. EFF, October 7, 2025
- 404 Media, May 27, 2025. "Immigration status" is exactly the category too: officers wrote "immigration" and "ICE" into the reason field, giving federal agents side-door access. 404 Media, May 27, 2025
The question: The ordinance covers surveillance technology, not pothole sensors. Name the "basic technology for everyday quality of life" that an approval-and-oversight requirement actually blocks, and tell me that cost is bigger than an unaudited ICE or abortion search.
No product has met the privacy standard he wants
[edit | edit source]license plate readers would have helped law enforcement officials track down and arrest the suspects. Systems like these need to prevent and solve serious crimes while protecting the privacy of community members, and I remain open to discussing the use of a resource that meets these standards.
written statement to Fox News Digital, May 19, 2026.
The record:
- Austin Police Department, May 13, 2025. Austin was already clearing serious crime at the ceiling without ALPR: "2024: 67/71 - 94% 2023: 75/75 - 100%," a 100% homicide clearance rate in 2023 before the pilot. Austin Police Department, May 13, 2025
- Virginia State Crime Commission, January 2026. The strongest possible safeguard, a binding statewide statute, failed within months. Virginia's law barred out-of-state and federal sharing, yet "13% (20 of 159) of agencies reported that law enforcement agencies in other states have continuous access" and "6% (9 of 159)" gave federal agencies continuous access. Virginia State Crime Commission, January 2026
- California State Auditor, February 2020. The documented-reason safeguard fails in practice: "In Sacramento, for 66 percent of searches, users left case number and search reason fields blank," and none of the audited agencies audited its own searches. California State Auditor, February 2020
- ACLU, July 2, 2026. The reason field is trivially defeated: one Oregon department searched Flock after typing "investigation" 111 times and "hehehe" 20 times, avoiding security-triggering words like "abortion." ACLU, July 2, 2026
The question: You say you are open to a resource that meets those privacy standards. Virginia's statewide statute, San Francisco's intelligence-center gate, and Flock's own reason field all failed that standard in documented practice. Name the one product or safeguard that has actually met it anywhere.