31 AUG 2026 — A YouGov survey of 20,563 American adults found 46 per cent opposed to automated licence plate readers in their community and 38 per cent in favour, a reversal from a year ago. The crossing is not the finding. Strong opposition went from 20 per cent to 31 per cent, and that is where almost the whole swing sits.

The two readings, side by side

YouGov asked the identical question in September 2025 and again on 24 August 2026, which makes this one of the rare year-on-year comparisons where the wording is not doing any of the work.

20% → 31%Strongly oppose, Sep 2025 to Aug 2026
17% → 14%Strongly support, over the same period
21% → 17%Not sure — the undecided block shrank
45% → 38% / 33% → 46%Total support and total opposition

Support fell seven points and opposition rose thirteen. Those do not net, because the missing four points came out of not sure.

Why the intensity number is the one to watch

A category that grows from 20 to 31 per cent has increased by more than half. This is the group whose opinion predicts behaviour.

Somewhat oppose is a survey answer. Strongly oppose is the bloc that shows up at a council meeting, files a records request, or organises a contract cancellation. It is now more than double strong support, at 31 per cent against 14. The 24 per cent who somewhat support the technology are not a constituency that will defend a procurement under pressure.

That asymmetry explains the reported pattern of dozens of communities rejecting or cancelling contracts better than the 46-to-38 topline does. A plurality against does not cancel anything. A committed minority against, facing indifferent support, does.

What the headline figure leaves out

Three caveats apply. None of them overturns the finding.

Opposition at 46 per cent is a plurality, not a majority. The 17 per cent who are unsure are larger than the eight-point gap between the two camps, so the crossing could be reversed by that group alone.

It is a single-day tracker question, fielded on 24 August, not a considered study with follow-ups. It tells you what people say when asked once, in one sentence, about a technology many will not have thought about before the question arrived.

And the question asks about use in your community for law enforcement purposes, which is the most sympathetic possible framing. A question naming retention periods, third-party data sharing or immigration enforcement would return different numbers, and YouGov asked a separate question about federal immigration use on the same day.

The same respondents, a different use, a different answer

YouGov put a second question to the same 20,563 people on the same day, and the pair is more informative than either alone.

Asked about the use of licence plate reader data by Immigration and Customs Enforcement to track undocumented immigrants, 40 per cent were strongly opposed and 20 per cent strongly in favour, with 50 per cent opposed overall against 36 per cent in favour. Strong opposition is nine points higher than on the general question, and strong support is six points higher too.

When a specific use is named, both poles harden and the unsure block falls from 17 to 14 per cent. Opinion has not settled at the level of the camera. It settles at the level of what the data is used for. A survey asking only about law enforcement in your community measures the softest version of the question.

It also explains why local fights over these contracts turn so quickly. A procurement is approved on the general framing and then argued about on the specific one.

Scale is what changed, not the technology

Automated plate readers are not new, and the objection is to the database rather than the camera.

Flock alone operates more than 120,000 readers across the United States. At that density the system stops being a set of cameras that catch a stolen car and becomes a searchable record of where vehicles have been, retained, queryable and shared between agencies. The privacy question attaches to the database and the retention policy, not to the optical component, which is why an argument about whether cameras deter crime tends to miss the point people are actually making.

The reported misuse cases matter here too. What moved this number is unlikely to be an abstract concern about surveillance; it is more likely the accumulation of specific reported instances of police using the network for purposes it was not sold for. People trust these systems on the basis of what the operator has been caught doing with them.

What this means where readers of this site live

The technology is regionally common, while the polling is not.

Plate recognition operates across ASEAN in car parks, tolling, enforcement cameras and estate access systems, generally without the consolidated national debate the United States is now having. Singapore's Personal Data Protection Act and Malaysia's PDPA both govern the resulting data, and neither is the constraint people assume it is: public-sector agencies sit largely outside Singapore's Act, and enforcement use is usually authorised by separate legislation rather than by consent.

The practical lesson from the American figures is about sequencing. Opposition rose after deployment reached scale and after misuse became public — the hardest moment to have the conversation, because the cameras are already up and the contracts are already signed. The same shape appeared in home surveillance, where Ring's new default encryption turned out not to be end-to-end by design — a defensible choice, disclosed late, arriving after the cameras were on the wall.