A county that votes 58 percent Republican rejected them by 58 percent
Knox County, Tennessee · likely voters
58.3%
oppose local police using Flock cameras. “Strongly oppose” alone, at 42.2 percent, is larger than every category of support combined.
Oppose 58.3%
Support 37.5%
This is not an uninformed reaction. 90 percent of respondents already knew about the program. Half, 50.1 percent, said they believe the cameras violate the Fourth Amendment, against 34.9 percent who said they do not. Among opponents, 73.9 percent gave government surveillance and privacy as their main reason, well ahead of cost or process. And the sample was 58 percent Republican, 53 percent conservative, and 64 percent aged 55 or older.
Method: Spry Strategies, LLC, 350 likely general-election voters in Knox County, TN, fielded August 15 to 17, 2026, text-to-web, margin of error ±5.2 points. Spry is a Knoxville-based polling and consulting firm and this is a non-probability opt-in sample, so treat the direction as reliable and the decimal points as approximate.
Keep the partisan composition in mind, because it kills the laziest response to this campaign. Opposition to warrantless mass plate surveillance is not a left-wing position. In Knox County it was the position of a Republican-leaning, older, highly engaged electorate that already knew what the program was.
Statewide in Arizona, 63 percent opposed
Arizona · registered voters
63%
oppose police use of automated plate readers in their community. 30 percent support it.
Other findings from the same survey:
- 82 percent are concerned about potential misuse of the data by law enforcement. 16 percent are not.
- 60 percent favour limiting or eliminating ALPR use, against 30 percent who want more of it.
- 82 percent support requiring independent audits and public reporting. That is the single most popular position in the poll, and it is exactly what Apex does not have.
- 51 percent oppose sharing local camera data with immigration enforcement such as ICE. 35.7 percent support it.
- 51 percent also say the cameras do help police solve certain crimes, such as stolen vehicles and hit-and-runs. People are not confused about the trade-off. They are rejecting it.
Method: NextGen Polling, 1,627 registered Arizona voters, fielded August 17 to 19, 2026, text-to-web, margin of error ±2.6 points, released August 20. NextGen is an Arizona-based political consulting and polling firm, and this is a non-probability opt-in sample.
A number we are not going to use, and why. Some coverage of this poll cited “59 percent support sharing data with ICE.” That is a Republicans-only subgroup figure. The full-sample result runs the other way: 51 percent opposed, 35.7 percent in favour. We are flagging it because we would rather tell you what a crosstab is than get caught leaning on one.
The finding that should worry a police chief
Peer-reviewed, randomized
Telling people what plate readers can do significantly lowered their trust in police.
Merola, Lum and Murphy ran a randomized survey experiment with 405 respondents. One group read brief, factual descriptions of what license plate recognition technology does. The control group did not. The group that learned about the technology reported significantly lower trust in law enforcement, and their strong agreement with positive statements about police, including that police respect individual rights, declined. Most respondents had no idea the technology was in use before being told.
This is the part that matters for Apex. The damage to public trust is not caused by activists, or by bad press, or by one officer misusing the system. It is caused by residents finding out what the system does. Every public information campaign defending this program is also, mechanically, the thing that erodes confidence in the department running it.
Citation: Merola, L. M., Lum, C., & Murphy, R. P., “The impact of license plate recognition technology (LPR) on trust in law enforcement: a survey-experiment,” Journal of Experimental Criminology, vol. 15, no. 1, pp. 55 to 66. Published online June 2018, print March 2019. This is an established study rather than a new one, and it is a single mid-Atlantic sample with high baseline trust in police, so it establishes causation on that sample rather than a national effect size.
The mechanism behind that result is well established in the procedural justice literature, which finds that people's willingness to accept police authority depends more on whether processes feel fair and transparent than on outcomes. That body of work is theory and correlation rather than a direct test of surveillance, but it is the accepted explanation for why a program adopted with no hearing and no vote costs a department legitimacy even when the technology works.
Federal cooperation is even less popular than the cameras
| Finding | Number |
|---|
| Disapprove of the job ICE is doing | 60 percent, against roughly 30 percent who approve |
| Say ICE's actions are making Americans less safe | 62 percent, against 37 percent who say more safe |
| Say ICE has “gone too far” | 65 percent, up from 54 percent in June 2025 |
| Partisan split on ICE disapproval | 91 percent of Democrats, 66 percent of independents. 73 percent of Republicans approve. |
Method: NPR, PBS News and Marist Poll, 1,462 US adults, fielded January 27 to 30, 2026, margin of error ±2.9 points, released February 5, 2026. All four figures come from this single national survey.
Regionally the picture is the same or worse. The University of Houston Hobby School surveyed 1,500 residents across nine Texas counties in April 2026, releasing results in May. 59.3 percent opposed cooperation between local law enforcement and federal immigration authorities, and 63.6 percent opposed their county sheriff working with ICE. In the same survey, 69.9 percent disapproved of how ICE is handling its job and 53.9 percent supported eliminating the agency entirely.
Put that against the record in Apex. APD's own Network Audit logs 9,700 searches by federal accounts reaching Apex cameras between January 2 and April 1, 2026, including 166 on January 29, the day the Chief told Council that Apex does not share with federal entities, hard stop. Roughly 3,300 searches of Apex cameras since January carry immigration-related reasons, run by outside agencies. The records are here.
Departments are backing out, and they are saying why
This is the part that should be in front of the Public Safety Committee. Chiefs and sheriffs are not abandoning these systems because activists asked. They are doing it because they can see what it is costing them.
| Where | What happened |
|---|
| Oshkosh, Wisconsin | Council renewed the Flock contract 5 to 2 on April 21, 2026. Less than 24 hours later it voted 7 to 0 to rescind, after the police chief said Flock representatives had misrepresented whether the system creates vehicle “heat maps.” Reported as the shortest gap on record between approving and cancelling a Flock contract. |
| Lynchburg, Virginia | On August 19, 2026 the police department paused all 15 of its fixed cameras pending an independent third-party review, stating it wanted to listen to community feedback and maintain public trust. The department was explicit that this was not prompted by misuse of its own system. |
| Henry County, Virginia | On August 14, 2026 the sheriff suspended all access to the Flock system effective immediately, citing privacy concerns, constitutional principles and the changing national landscape, and began dismantling 14 cameras the following week. |
| Henry County, Illinois | On August 20, 2026 the sheriff's office announced it would end its 8-camera program effective August 31. |
| Knox County, Tennessee | The sheriff's office paused its plan for nearly 300 cameras in August 2026 and then opted out, the same month the poll above was published. |
| The Triangle | Chatham County terminated in April 2026. Pittsboro voted to end its programme in May with a July 1 removal deadline. Hillsborough dropped an $81,500 contract over data-sharing language. More than 80 localities nationwide had dropped Flock by May 2026. |
Every one of those is a department that concluded the reputational cost was higher than the investigative benefit. Apex is being asked to reach the opposite conclusion on the same evidence, at the same moment, with 9,700 federal searches in its own audit log.
Why the public reacts this way: the documented record
People are not reacting to a hypothetical. They are reacting to reporting.
- Electronic Frontier Foundation, December 2025. Analysis of datasets covering more than 12 million searches logged by more than 3,900 agencies between December 2024 and October 2025. Findings included hundreds of searches tied to protest activity, audit entries containing ethnic slurs against Roma people, and a Johnson County, Texas deputy who logged the search reason “had an abortion, search for female,” a query that reached 6,809 networks and 83,345 cameras.
- 404 Media, from May 2025. Records from Danville, Illinois showed more than 4,000 nationwide and statewide lookups with immigration-related reasons between June 2024 and May 2025, in a state whose law bars ALPR use for immigration enforcement. Later reporting documented Customs and Border Protection with access to more than 80,000 cameras, and Senator Wyden's office documented searches by ICE's Homeland Security Investigations division, the Secret Service and Navy criminal investigators. Two members of Congress opened a formal investigation into Flock as a result.
- Brookings Institution, May 20, 2026. Stephanie Pell argues that federal acquisition of surveillance technology combined with Congress's failure to regulate it accelerates the loss of public trust in the Department of Homeland Security. Worth reading for that thesis. Its central case study is cell-site simulators rather than plate readers, so we cite it for the trust argument and nothing more.
And the number nobody has: Apex
No scientific survey of Apex residents on this program has ever been commissioned, by the Town, by the department, or by us. Council has never measured local opinion on a surveillance system it has never voted on. That is its own finding.
What the record does contain: on August 11, 2026 the Council chamber filled and the town opened overflow seating on the third floor. Eight residents spoke against surveillance technology. One spoke in favour. Council pulled a four-year Axon technology amendment off the consent agenda before public comment began.
We run a poll on our own front page. It is a self-selected click poll on an advocacy site, it is not evidence, and we are not going to pretend otherwise. If the Town wants a defensible number for Apex, it can commission one before it decides whether to renew. We would publish the result either way.
How to read this page. Two of the polls above, Knox County and Arizona, come from partisan or commercial firms using non-probability text-to-web samples. Their direction is consistent with everything else on this page, but they are not gold-standard probability samples and we are not going to dress them up as such. The Harris Poll finding that 74 percent of Americans think municipalities should be able to switch specific AI camera features on and off was commissioned by LiveView Technologies, a security-camera vendor, and should be read as sponsor-funded research. The investigative work by EFF and 404 Media documents real misuse. It does not by itself measure a population-level fall in trust, which is why the randomized experiment above matters more than any of it.
Sources
- Spry Strategies, Knox County Flock Cameras: Survey Results, fielded August 15 to 17, 2026.
- NextGen Polling, Arizona statewide ALPR survey, fielded August 17 to 19, 2026, released August 20. Covered by AZFamily/KOLD, 12News, The Epoch Times and The Blaze.
- NPR, PBS News and Marist Poll, The Actions of ICE, fielded January 27 to 30, 2026.
- University of Houston Hobby School, Center for Public Policy Report 10-2026, fielded April 2026.
- Merola, Lum & Murphy, The impact of license plate recognition technology on trust in law enforcement, Journal of Experimental Criminology 15(1), 2019.
- EFF, Investigations Expose Flock Safety's Surveillance Abuses: 2025 in Review, and the Johnson County, Texas abortion-search reporting.
- 404 Media, ICE Taps into Nationwide AI-Enabled Camera Network, and subsequent reporting on CBP access and the congressional investigation.
- Stephanie K. Pell, Brookings Institution, May 20, 2026.
- Harris Poll for LiveView Technologies, 2026 Public Safety and Privacy Benchmark, fielded March 23 to 25, 2026, released June 1. Vendor-sponsored.
- Department reversals: WBAY, Fox 11 and NBC26 on Oshkosh; WDBJ, WSLS, WSET and Cardinal News on Lynchburg; WDBJ, WSLS and the Martinsville Bulletin on Henry County, Virginia; WVLT and Knox Pages on Knox County, Tennessee.