The Chilling Effect Is the Feature, Not the Bug

There’s a line buried in the San Francisco audit story that stopped me mid-scroll. It notes that civil liberties groups argue people may avoid attending protests, seeking reproductive health care, or participating in political activities if they believe their movements are being tracked.

One commenter on the thread put it plainly: that is entirely the point.

I think they’re right. And that’s what makes this particular technology conversation different from most of the noise around privacy. We’re not talking about a data breach, or a company selling your purchase history to advertisers. We’re talking about a system where the chilling effect on lawful behaviour is arguably the intended outcome, whether anyone says so out loud or not.

The Flock license plate camera network is American, but the architecture of the problem is not. Victoria has had its own rolling debates about automated number plate recognition, about who holds the data, for how long, and who gets to query it. The answers have never been fully satisfying. They tend to arrive in the form of policy documents that assume good faith from every future operator of the system, which is a generous assumption to build critical infrastructure on.

What the San Francisco audit found is predictable in hindsight: outside agencies accessed the data, the policies were thin, the oversight was thinner. This is almost always how it goes. The system gets built under one set of assumptions, then the access creeps quietly outward. Not always maliciously. Sometimes it’s just an underfunded agency taking a shortcut because the shortcut is right there.

The technical framing tends to dominate these discussions, which suits the vendors. Talk about data partitioning and access controls long enough and you’ve successfully moved the conversation away from the prior question: should this exist at all. I’ve watched that move happen in enough enterprise security contexts to recognise it. Complexity is a great place to hide a values question.

Some people in the thread were talking about physical resistance to the cameras, which I understand as frustration even if I wouldn’t go there myself. The more grounded suggestion was showing up to council meetings. Someone pushed back hard on that, pointing out that local councils tend to defer to police departments on this stuff, and the police departments love these systems. That tension is real and I don’t have a clean resolution for it. Democratic process is slow and it tends to lose to incumbent institutional interests, but it’s also the only tool that builds durable change rather than just replacing one camera that gets knocked over.

The thing that genuinely unsettles me is the self-censorship angle. If someone decides not to attend a protest because they don’t know who will query that plate data later, or under what future government, then the system has already done its work without anyone needing to act on the data at all. The database doesn’t need to be misused. Its existence is enough.

That’s not a hypothetical concern. The landscape of what constitutes a prosecutable political act can change faster than the data retention window on one of these systems. Australia is not immune to that drift.

I don’t know where the equilibrium is between legitimate public safety uses of this kind of technology and the cumulative effect of building a movement-tracking infrastructure into every suburb. I’m not sure anyone does. But I’m increasingly suspicious of frameworks that treat that as a technical problem with a technical solution, rather than a political problem that requires a political answer.