When I was a college student, I asked a question in a political science class that seemed straightforward at the time.

Why did Soviet citizens put up with it?

Why did they accept government surveillance, restrictions on speech and the absence of freedoms that Americans considered fundamental?

My professor’s answer has stayed with me for decades. He suggested that people who have never experienced a different reality often don’t recognize what they are missing. Most people are focused on work, family and the practical challenges of everyday life. Liberty, privacy and civil rights can feel abstract until they begin to disappear.

At the time, I viewed that explanation as something relevant to another place and another era. America was different. Whatever our flaws, we were a country that prided itself on limiting government power and protecting individual freedoms.

Lately, I have found myself thinking about that conversation again.

A recent article by Jon Swartz on Digital CxO examined reports that federal intelligence and law enforcement agencies are monitoring categories that reportedly include anti-technology extremists, anti-Christian activity and opposition to certain government policies. Whether every aspect of the reporting ultimately proves accurate is less important than the broader issue it raises. The story is not simply about surveillance. It is about surveillance in an era when technology is fundamentally changing what surveillance means.

For most of modern history, governments faced practical limitations. Information had to be collected, organized, reviewed and acted upon by human beings. Surveillance required investigators, analysts, informants, files, budgets and time. The process was intrusive, but it was also constrained by economics and manpower.

Technology is removing many of those constraints.

Artificial intelligence can process volumes of information that would have overwhelmed entire agencies a generation ago. Computer vision systems can identify individuals across massive collections of images and video. Data brokers compile information about where people travel, what they purchase and increasingly how they behave online. Agentic systems are being developed to monitor environments, identify anomalies and generate recommendations with minimal human involvement.

Most discussions about AI focus on business applications. We hear about autonomous enterprises, digital workers and intelligent agents capable of automating workflows. The technology industry understandably views these developments as opportunities to improve productivity, reduce costs and accelerate decision-making.

The same capabilities, however, have implications that extend well beyond the enterprise.

An AI system does not particularly care whether it is monitoring supply chains, customer service interactions or human behavior. The underlying mechanics are remarkably similar. Data is collected. Patterns are identified. Relationships are mapped. Conclusions are generated.

That reality creates questions that receive far less attention than they deserve.

The technology sector spends considerable time discussing responsible AI, model alignment and governance. Most of those conversations focus on the behavior of the models themselves. We worry about hallucinations, bias, explainability and reliability. Those concerns are legitimate, but they address only part of the challenge.

The larger question may be how institutions choose to use increasingly powerful systems once those systems become available.

History offers plenty of examples of surveillance authorities expanding beyond their original purpose. Governments change. Political priorities evolve. New threats emerge. Yet surveillance capabilities, once established, tend to become permanent features of the landscape. The technologies improve. The databases grow. The scope broadens.

That observation should concern people across the political spectrum because it is not fundamentally a partisan issue. The individuals operating a system today are rarely the same individuals who will operate it tomorrow. The capabilities being developed now will almost certainly outlast the political debates that surround them.

The distinction between monitoring behavior and monitoring beliefs becomes especially important in this context. Investigating criminal activity is one thing. Monitoring lawful dissent, unpopular opinions or criticism of institutions is something entirely different.

Democratic societies depend on the ability of citizens to disagree with one another and with their government. The health of a free society is often measured not by how it treats popular viewpoints, but by how it treats unpopular ones.

Technology is making that distinction harder to ignore because it is steadily reducing the cost of observation.

Previous generations worried about governments collecting too much information because collecting information was difficult. Today, the challenge is different. Information is abundant. Processing power is abundant. Storage is abundant. The barriers that once limited large-scale surveillance are disappearing.

Many of the technologies enabling this transformation were created with entirely legitimate objectives. Facial recognition can improve security. Data analytics can help detect fraud. AI systems can identify cyber threats more quickly than human analysts. None of these technologies is inherently malicious.

That is precisely what makes the issue complicated.

The risk does not arise from a single technology. It emerges from the accumulation of capabilities and the gradual normalization of their use. Each new tool can be justified on its own merits. Each deployment appears reasonable in isolation. Over time, however, the combined effect can produce something far more powerful than anyone originally envisioned.

The technology industry often talks about trust as a prerequisite for AI adoption. Customers must trust the systems. Employees must trust the systems. Regulators must trust the systems. That conversation usually focuses on whether AI produces accurate outcomes.

Trust is also shaped by how technology affects the relationship between citizens, institutions and power.

If people begin to believe that every online activity, every association and every expression of disagreement is potentially subject to analysis, categorization and permanent retention, public attitudes toward both technology and government will inevitably change. Suspicion has a way of spreading. Once it takes hold, rebuilding confidence becomes difficult.

The challenge facing society is not whether artificial intelligence should be developed. That debate is already over. AI will continue to advance because the economic incentives are too powerful and the potential benefits are too significant.

The more interesting question is whether the safeguards surrounding these technologies are evolving as quickly as the technologies themselves.

That is ultimately what concerned me about the Swartz article.

Not the politics of the moment.

Not the personalities involved.

The realization that we are entering an era in which surveillance is becoming increasingly automated, increasingly scalable and increasingly inexpensive. Previous generations worried about governments knowing too much because acquiring information was hard. Our generation is confronting the opposite problem.

The cost of observation is approaching zero.

Shimmy’s Take

For decades, limitations in technology acted as a natural check on surveillance. Governments could only monitor so much because collecting, organizing and analyzing information required enormous effort.

Artificial intelligence is changing that equation.

The capabilities now being built for enterprises, security operations and digital transformation projects are extraordinarily powerful. Many will deliver tremendous benefits. Some already are.

But history suggests that whenever a new capability emerges, someone eventually finds a way to use it beyond its original purpose.

The debate is no longer about whether large-scale surveillance is technically possible. It is.

The first question is whether our laws, institutions and expectations about privacy are keeping pace with the technologies that are reshaping them. The ultimate question goes back to what my professor told me all those years ago. We know what it is like to have real freedom; we are not half-starved, only concerned about our next meal. We know better. How long are we going to stand for this?