Have you ever driven past a slim box on a pole and felt, just for a second, that the street was looking back? I have. It is a small, almost silly thought until you realize how many of those boxes now talk to one another. This week that uneasy feeling moved from the curb to a hearing room. Four chief executives who build and sell AI surveillance cameras were invited to explain what their networks actually do. None of them sat in the chair.
Empty Chairs And A Growing Camera Web
The invitation was not a casual coffee chat. The chair of a Senate crime subcommittee asked leaders from Flock Safety, Axon, Motorola Solutions, and Verkada to walk lawmakers through capabilities, customers, and safeguards. Ranking members on the other side of the aisle signed on as well. Before the session, there was even talk that a no-show might eventually trigger a subpoena. Still, the seats stayed empty.
I do not think that means the companies are cartoon villains. Running a large public-safety business is messy. Legal teams write careful letters. Calendars collide. But optics matter, and this particular optic was rough. Senators wanted a plain-language account of a market that now stitches thousands of cameras into searchable pools of vehicle images. What they got instead was a hearing about absence.
We need to hear from them about what it is their networks do. We need to hear from them about their intentions, about their capabilities.
That is the heart of it. Not a morality play. A request for a map. How big is the web? Who can query it? What happens when a hit is wrong? Those are ordinary questions. They become urgent once you hear how cheap entry can be.
How A Single Camera Becomes A National Search
Here is the part that sticks with me. A customer does not need a citywide grid on day one. Buy one unit. Pay a subscription that, according to remarks at the hearing, can land near a few thousand dollars a year. In return, that customer may tap a much larger pool. In one firm’s case, lawmakers cited on the order of 120,000 cameras and billions of stored images.
Think of it as a library card. You put one book on the shelf. You get to browse the whole wing. That is efficient for a detective chasing a stolen van. It is also the feature that makes civil-liberties people sit up straight. Scale changes the meaning of a tool. A camera on one alley is a camera. A searchable nation of plates is a system.
One company pushed back on the “national database” label. A spokesperson said no customer can search a nationwide Verkada network because none exists, and that users only see data inside their own systems. Fair enough to put that on the record. Product designs differ. Some firms emphasize shared lookup. Others insist on local silos. The public debate often lumps them together, which is sloppy. Precision helps.
- Entry cost can be a single camera plus an annual plan
- Shared networks can expose far more imagery than one town owns
- Not every vendor offers the same search model
- Misuse and error are no longer theoretical talking points
Why Plate Readers Feel Different From A Store Camera
A shop cam watches a doorway. An automated license plate reader watches motion through public space and turns that motion into structured data. Time. Place. Plate. Direction. Repeat that millions of times and you get a timeline of ordinary life: school drop-off, night shift, a weekend out of town.
I’ve found that people accept cameras more readily when the story is a recovered car or a missing person. Those stories are real. Agencies credit plate networks with closing cases that would have gone cold. That is not marketing fluff in every instance. Detectives will tell you a hit can be the difference between a lead and a shrug.
The other side of the ledger showed up in the same hearing room. A woman described being wrongly tied to a fatal crash after a camera error. Separately, five officers in Indiana faced charges tied to alleged misuse of the technology. You can support lawful policing and still wince at those facts. Tools do not stay virtuous just because the brochure says “public safety.”
A customer in a Flock network, for example, just needs to buy a single camera and a subscription… and in return they get access to a nationwide database of tens of thousands of cameras.
Read that slowly. The sentence is not poetry. It is architecture. Architecture decides whether a search is local curiosity or a cross-country trawl. Architecture also decides who logs the query, who audits it, and who gets fired when someone looks up an ex instead of a suspect.
What The Companies Say When They Do Speak
Absence at the table does not mean total silence. Flock’s chief legal officer sent a letter dated September 8 describing work over the past year to prevent and detect misuse. The company also said it would support “clear, sensible legislation” because durable rules sit under public trust, and public trust sits under the business.
That last clause is worth circling. It is unusually candid. If your product lives or dies on police contracts and city councils, you cannot afford a reputation as the firm that shrugs at abuse. Supporting legislation can be sincere. It can also be a way to shape the rule before someone else writes a harsher one. Both can be true at once. Adults in this market know that.
Verkada, for its part, said its chief executive could not attend, that the firm was already sharing information with the subcommittee, and again that customers cannot roam a national Verkada search because that network is not a thing. Other invited firms did not offer public comment in the immediate aftermath. That vacuum filled itself with suspicion. Maybe that is unfair. Vacuums still fill.
Bipartisan Irritation Is Not The Same As A Ban
One striking detail: the skepticism did not split neatly by party. That matters. Tech fights in Washington often become team sports. This one sounded more like a shared headache. Members wanted executives on camera, almost literally, to answer basic capability questions. A ranking senator even joked that Flock leaders seemed nervous about being on camera. Cheap line. Memorable line.
Irritation is not a statute. No one at the session flipped a switch that turns off pole cameras tomorrow. What you heard instead was a demand for an “honest conversation about where this is leading us as a country.” That phrase is vague until you translate it. Translation: retention limits, query logs, purpose restrictions, penalties that bite, and maybe a brighter line between a city system and a national lookup.
Congressional committees can compel testimony. The chair noted they had not issued subpoenas yet and suggested they should. Whether that happens is politics plus timing plus lawyers. Investors in listed names tied to this sector should treat that as a process risk, not a punchline.
The Public Safety Case Nobody Should Hand-Wave Away
It would be lazy to write this as a horror story only. Plate readers help recover vehicles. They help reconstruct routes after a shooting. They can place a suspect’s car near a scene when witnesses freeze. For small departments, buying into a shared network can feel like borrowing a big-city tool without a big-city budget.
In my experience covering tools like this, the most convincing advocates are not marketers. They are investigators who remember a specific night when a hit arrived in time. Those stories should stay in the file. Ban-first rhetoric that ignores solved cases will not survive contact with victims’ families. The honest fight is over guardrails, not over whether crime exists.
Still, “it solves crime” cannot be the last sentence. Plenty of powerful systems solve problems and still need limits. Bank records solve fraud. We do not hand every clerk the whole ledger. Phone location can find a lost hiker. We still argue about warrants. Cameras sit in that same family now, whether vendors like the comparison or not.
False Hits, Curiosity Searches, And Human Weather
Software reads plates. Humans read motives. That gap is where trouble lives. A misread character on a plate can send police to the wrong driveway. A jealous search can turn a workplace tool into a stalking aid. A sloppy policy can let a login wander for months without review.
The Indiana charging news landed on the same day as the hearing. Coincidence of calendar, maybe. It still framed the discussion. When officers are accused of abusing a lookup tool, the product is no longer an abstraction. It is an evidence trail with names on it.
Lindsey Isaacs told the panel she was wrongly pulled into a fatal-crash theory because of a camera mistake. Sit with that. Even if the error rate is low in percentage terms, the personal cost of one bad match is enormous. Low error plus huge scale still produces a pile of wrecked afternoons. Scale is the multiplier people forget.
- Define allowed purposes in writing, not in hallway talk
- Log every query with a reason code that a supervisor can audit
- Set retention so old trips do not live forever by default
- Create a fast path to correct a bad hit and notify the person harmed
- Punish curiosity searches hard enough that gossip is expensive
None of those steps require hating cameras. They require treating the network like a sensitive database, because that is what it is.
Markets, Letters, And The Rumor Of A Sale
Two of the invited firms are familiar tickers in public markets. Hearings rarely move a chart by themselves. They do change the tone around diligence. Contract risk, legislative risk, and headline risk start showing up in the same paragraph. That is already happening.
There was also market chatter that Flock has explored a sale. Treat rumors as rumors. What is not a rumor is the letter promising better misuse controls and openness to legislation. A company that might want a clean exit, or a richer valuation, has extra reason to look governable. Buyers hate surprise subpoenas. Cities hate surprise scandals. Those incentives can align with reform, at least on paper.
| Pressure Point | Why It Matters | Near-Term Signal |
| Empty testimony chairs | Raises odds of compulsory process | Watch for subpoena language |
| Shared lookup design | Turns one camera into wide search | Watch contract language on access |
| Documented misuse | Moves debate from theory to cases | Watch local charging and audits |
| Support for legislation | Vendors try to shape the rulebook | Watch draft bills on retention |
If you hold the stocks, this is not a day-trading headline. It is a governance overlay on a growth story. Growth stories in public safety can be sturdy. They get less sturdy when councils pause renewals or when statehouses start drafting purpose limits.
What “Clear, Sensible Legislation” Could Actually Mean
Everyone loves the adjective sensible until someone writes a definition. So let’s try a concrete sketch, the kind a city attorney could annotate.
First, purpose. Lookups should require a case number or an articulable public-safety reason, not a vibe. Second, sharing. If a small town taps a giant pool, the town should accept the same audit standard as the pool’s biggest user. Third, retention. Keeping every pass-by for years is convenient for later investigations and corrosive for ordinary movement. Fourth, notice. People who were wrongly flagged deserve a correction path that does not require a lawyer on speed dial.
Fifth, and this one is unfashionable, vendor liability when defaults invite abuse. If a product makes gossip easy and audit hard, the design is part of the harm. I am not arguing for reckless lawsuits. I am arguing that “user error” is a weak shield when the user is given a nationwide search box and a thin policy PDF.
A rough balance many cities want: Solve serious crimes faster Keep everyday trips from becoming a permanent dossier Make misuse painful, documented, and public enough to deter
Perhaps the most interesting aspect is how quickly this stopped being a niche procurement topic. Homeowner associations, school lots, retail parks, and police departments can sit on the same map. That mix is why the hearing felt larger than one brand.
Living With Cameras Without Going Numb
Most of us will not move to a cabin to escape poles. We will keep driving to work. We will keep parking at the grocery store. The adult move is not panic. It is literacy. Ask your city who can search. Ask how long images live. Ask whether a neighbor association’s camera feeds a police pool. Those are boring questions. Boring questions prevent ugly surprises.
I keep coming back to the empty chairs. Showing up would not have ended the controversy. It might have narrowed it. Executives could have drawn lines between products, explained audit logs in plain English, and said out loud what they will not build. Instead, lawmakers narrated the network for them. That is rarely the version a company prefers.
Will there be subpoenas? Maybe. Should there be a single federal template for plate-reader use? Debatable. Can a country keep the crime-fighting upside and still refuse a casual, always-on travel diary? That is the live question, and it does not care whether a CEO had a conflict on the calendar.
We need to have an honest conversation about where this is leading us as a country.
Honest conversations require the people who built the pipes. Until they sit down, the public will fill in the blanks, and the blanks are not flattering. Cameras will keep going up either way. The only open issue is whether the rules grow up at the same speed as the lenses.
A Practical Checklist For Readers Who Want More Than Outrage
Outrage is easy and expires fast. If this topic actually bothers you, do smaller things that leave a paper trail.
- Read your city’s camera contract, especially sharing and retention clauses
- Ask whether hot-list alerts need a supervisor sign-off
- Request audit summaries, even redacted ones
- Push for a public annual report on query volume and disciplinary cases
- Separate emergency pings from fishing expeditions in policy text
None of that makes you anti-police. It makes you a grown-up customer of government. Cities buy these systems with public money. Public money comes with public questions. Vendors who want long contracts should want those questions answered in daylight.
And if you work in the industry, a piece of unsolicited advice from someone who has watched too many hearings: send the CEO. Send the person who can say “here is what the software cannot do” without glancing at a lawyer every ten seconds. Technical humility plays better than a polished no-comment. The market for trust is thinner than the market for cameras.
Where The Story Goes After The Gavel
Hearings are theater with footnotes. The footnotes will matter more than the theater. Watch three threads. First, whether compulsory testimony arrives. Second, whether state legislatures copy each other’s retention bills. Third, whether a high-profile false hit or abuse case lands in a jurisdiction that already disliked the cameras.
Any one of those can slow deployments without a national ban. Slow is not stop. The installed base is already large. Taking cameras down is harder than putting rules on new ones. That is why the next six months of contract language may shape the next six years of street life.
I started this with a pole on a wet road. I’ll end there too. The box does not care about your errands. It cares about a string of characters and a timestamp. We are the ones who decide whether that string becomes a tool, a trap, or both. The executives who skipped the room still have to live with that decision. So do the rest of us, every time the light turns green and the lens stays on.
