Being able to do it
is not permission.

Something interesting happened last week where I live in Volusia County, Florida.

Sheriff Mike Chitwood announced that the Volusia Sheriff's Office was shutting down its license plate readers.

And here's what makes this debate worth having:

The cameras worked.

According to Sheriff Chitwood, license plate readers helped Volusia County solve violent crimes, find missing people, recover stolen property and arrest murderers, armed robbers, burglars, hit-and-run drivers, sex offenders and child predators.

Just a week before announcing their discontinuation, the Sheriff's Office reported using Flock to identify the tag of a homicide suspect.

So let's not pretend this is an easy argument.

There are real public-safety benefits to this technology.

But something else is happening.

Florida has now revoked permits for automated license plate readers in state-highway rights-of-way and stopped issuing new ones.

Governor Ron DeSantis has said their proliferation had gotten “out of control.”

And Florida isn't alone.

Flock Safety has built a network of roughly 120,000 cameras across 49 states. According to AP, the technology is being used in approximately 6,000 communities.

But public sentiment appears to be shifting.

A recent Reuters/Ipsos poll found 47% of Americans opposed their community using Flock cameras, compared with 38% who supported it.

More than 50 jurisdictions have reportedly ended their Flock contracts.

Flock itself has responded to the controversy by announcing additional safeguards, including shorter default retention periods, case-number requirements and tools designed to identify abnormal searches.

And opposition is coming from both sides of the political spectrum.

That should tell us something.

Maybe the debate isn't really about license plates.

Maybe it's about surveillance.

Useful doesn't necessarily mean acceptable.

That's the uncomfortable part.

If we put enough cameras everywhere, connect enough databases and give increasingly powerful AI access to the information, we will absolutely discover useful things.

We will solve crimes.

We will find missing people.

We will identify dangerous behavior.

We will probably prevent some tragedies.

Take the principle far enough and we could build an extraordinarily safe society.

We'd just have to watch everyone.

Everywhere.

All the time.

And remember what we saw.

Most people instinctively recognize that something is wrong with that bargain.

There is a word ordinary people often use when technology crosses that invisible boundary between helping us and watching us:

Creepy.

Engineers probably won't like that word.

Lawyers can't easily define it.

Marketers definitely don't want it appearing in the brochure.

But human beings understand it remarkably well.

And perhaps we should pay more attention to that instinct.

Now let's move the camera.

Take it off the highway.

Put it inside your home.

More specifically, put it in your bedroom.

Now imagine that you're 84 years old.

The camera isn't there because you've committed a crime.

It's there because someone wants to keep you safe.

Maybe it can detect whether you've fallen.

Maybe AI can determine whether your behavior has changed.

Maybe someone can reconstruct exactly what happened before an incident.

Maybe the system can determine who entered your room, when they arrived, how long they stayed and what happened while they were there.

There are legitimate uses for all of that information.

Sound familiar?

This is increasingly part of the technology conversation happening in senior living.

And I think our industry needs to pay very close attention to what's happening with Flock.

Because society may be beginning to draw a boundary that technology companies have been racing past:

There is a difference between detecting a specific event and continuously observing a human being because the resulting information might eventually prove useful.

Older adults don't stop valuing privacy because they need help.

There is research behind this.

A 2024 study specifically examined older adults' attitudes toward cameras used for fall detection.

Researchers surveyed 389 older adults, conducted interviews and then involved older adults in a privacy-focused design exercise.

They found something important.

Privacy concerns weren't merely about someone stealing data.

They were connected to dignity and control.

Earlier research examining cameras inside nursing-home rooms warned about threats to resident privacy and dignity and the possibility that technology installed for one purpose could expand into broader monitoring of residents and staff.

That last part deserves particular attention.

Because technology almost never becomes less capable.

AI gets better.

Computer vision gets better.

Storage gets cheaper.

Models become capable of making increasingly sophisticated inferences.

Databases become easier to connect.

What begins as:

“Did Malcolm fall?”

can eventually become:

How often does Malcolm get out of bed?

How well is Malcolm walking?

Who visited Malcolm?

How long did they stay?

How quickly did a caregiver respond?

What did the caregiver do?

Has Malcolm's routine changed?

Is Malcolm becoming more confused?

Is Malcolm receiving more care than he's paying for?

What else can we infer?

Some of those questions might be incredibly valuable.

That doesn't automatically give us the moral authority to collect the information required to answer them.

This is where I think technologists sometimes get the question backward.

If a resident consents to fall detection, our engineering challenge shouldn't be:

How much useful information can we collect now that we have a camera in the room?

It should be:

What is the least information we need to observe to reliably determine that this person has fallen?

Those are profoundly different design philosophies.

One starts with surveillance and tries to bolt privacy controls onto it afterward.

The other starts with human dignity and deliberately limits what the technology is capable of knowing.

I'm increasingly convinced that the second approach is where responsible AI needs to go.

Not because cameras are inherently evil.

Not because public safety doesn't matter.

And certainly not because technology shouldn't help protect vulnerable people.

Quite the opposite.

Our technology should be good enough that protecting someone doesn't require unnecessarily watching them.

And this brings me back home to Volusia County.

I respect Sheriff Chitwood's frustration.

He can point to actual criminals who were caught because this technology existed.

That is a serious argument.

But it also creates a question I don't think we should run away from:

How much privacy would you personally surrender in exchange for additional safety?

Would you allow your movements to be recorded whenever you drive?

Would you allow those observations to be searchable?

Would you allow different organizations to share them?

Would your answer change if the data were retained for a month?

A year?

Forever?

Would you accept cameras throughout your own home if someone could demonstrate that doing so statistically reduced your risk of injury?

Your bedroom?

Your mother's bedroom?

At what point would you finally say:

That's creepy.

I don't think there is one simple answer.

But I increasingly believe there is one principle worth defending:

Solve the specific problem a person has consented to solve using the least invasive technology reasonably capable of solving it.

Detect the fall.

Find the missing resident.

Send the emergency call.

Protect the person.

But don't turn someone's life into a dataset simply because someday the data might be useful.

Privacy makes some problems harder to solve.

Liberty does too.

That's part of the price of living in a society where human beings are more important than optimization.

The technology industry is going to become extraordinarily good at watching people.

I think the more important innovation will be learning how not to.

What do you think?

Where is your line between safety and surveillance?

And would that line change if the person being watched were you?