The comfortable consensus in Washington holds that government surveillance infrastructure is a necessary evil. Democrats worry about abuse. Republicans worry about inefficiency. Both camps assume the machinery will persist, debated at margins but fundamentally accepted.
This consensus is about to shatter. Not because activists finally win the argument, but because the infrastructure itself is losing the one thing that kept it politically stable: the appearance of competence.
Recent reporting on how algorithmic tools made their way into the NHS, and earlier stories about ICE's operational failures despite expanded surveillance powers, point to a deeper fracture. We've built an enormous surveillance apparatus on the assumption that it would work. Now we're learning it doesn't work very well. Worse, we're learning this the hard way, in real time, when people's lives hang in the balance.
The old political equilibrium depended on surveillance being invisible and presumed effective. You don't see the flight you almost took because an algorithm flagged something suspicious. You don't know the person detained for three days on a data error. The system operates in shadow, so its failures stay hidden too. As long as that opacity held, the debate stayed abstract. Surveillance advocates could claim it prevents catastrophes. Critics could warn about dystopia. Neither side could prove their case.
That's changing.
When a government agency uses surveillance tools to detain the wrong person, or when predictive systems get embedded in critical infrastructure like healthcare and produce worse outcomes, the failure becomes visible and measurable. It's not a philosophical dispute anymore. It's a concrete problem: the system we built to make us safer is sometimes making things less safe.
This creates a political opening that neither the surveillance enthusiasts nor the privacy maximalists fully anticipated. The question isn't whether we should surveil. It's whether we should keep paying for surveillance that demonstrably doesn't deliver.
Consider what this could break. The current coalition supporting surveillance expansion relies on a quiet agreement: Democrats get to express concern about fairness, Republicans get to express confidence in security, and both get the surveillance state they want. Everyone's happy because nobody has to admit it doesn't work very well.
The moment we start seriously measuring outcomes, that coalition fragments. Why fund an immigration system with expanded surveillance powers if it still produces mass errors? Why embed algorithmic tools in healthcare if they don't improve diagnosis? Why expand any system if we can't show it actually solves the problem it was supposed to solve?
Technology companies and government agencies have a built-in defense: they can always claim the system needs more data, better training, or different implementation. That works for a few cycles. But eventually the credibility tax compounds. Each failed promise makes the next argument harder to sell.
What comes next is harder to predict. We might see genuine privacy reform, not from moral arguments winning but from fiscal arguments: can we afford this? We might see a chaotic fragmentation, where different agencies and companies pursue incompatible approaches. Or we might see a shift toward smaller-scale, more targeted surveillance accepted precisely because we've given up on the comprehensive approach.
The real political disruption isn't coming from privacy advocates finally making their case. It's coming from surveillance systems failing to deliver their core promise, then having nowhere to hide because their failures are now matters of public record.
The consensus that accepts surveillance as necessary was always fragile. It depended on people not looking too closely at whether it actually works. That dependence is ending.