MIT Is Spending $3 Million to Turn Its Campus Into a Surveillance Lab. Nobody Should Be Okay With This.

You’d think the institution that gave the world the free software movement would know better. You’d think a place built on the radical idea that knowledge should be open, shared, and free might pause before blanketing its campus with 500 AI-powered cameras designed to watch, categorize, and file away every person who walks by.

You’d be wrong.

MIT is spending over $3 million on a surveillance system that doesn’t just record — it classifies. Clothing color. Gender. Age. Up to 35 feet away. Automatically. Without your knowledge. Without your consent. Without a conversation about whether this is the kind of place anyone actually wants to be.

When an institution starts sorting people by what they wear and how old they look, it’s not building safety. It’s building a database. And databases don’t forget.

Let’s be clear about what’s happening here. These aren’t your grandfather’s CCTV cameras — grainy footage pulling to a monitor nobody watches. These are algorithmic classification systems. They see you. They tag you. They reduce you to a row in a table: male, 20s, blue jacket, 2:47 PM, Building 10 entrance. And they do this thousands of times a day, building a record of movement, pattern, and behavior that becomes the training data for something much larger.

The comment that should keep you up at night is the obvious one: Why not skin color at this point? Because once you’ve accepted that a camera should classify people by gender and age, you’ve already accepted the premise. You’ve already agreed that institutions have the right to sort human beings into categories in real time. Skin color is just another column in the same database. The line you thought was a wall was always a suggestion.

The most dangerous surveillance isn’t the kind that catches criminals. It’s the kind that teaches a society to stop noticing it’s being watched.

MIT will tell you this is about safety. They’ll point to incidents, to concerns, to the language of risk management and institutional responsibility. And some of that might even be sincere. But safety is the universal solvent for civil liberties — pour it over any right, and watch it dissolve. The logic is always the same: something bad happened, therefore we must see everything, therefore you must be seen.

But here’s the paradox they won’t confront: a campus that monitors your clothing color and estimates your age from 35 feet away doesn’t feel safe. It feels like a panopticon. It feels like a place where the walls have eyes and the eyes have opinions about who you are. The very surveillance deployed to create security destroys the conditions — autonomy, privacy, freedom from scrutiny — that make a place feel secure in the first place.

You cannot engineer trust with a camera. You can only engineer compliance. And a campus that chooses compliance over trust has already lost the thing it claims to protect.

What makes this especially galling is the location. This is MIT. This is the institution of Aaron Swartz, who fought against the enclosure of knowledge. This is a place that has produced some of the most important thinking on privacy, on algorithmic ethics, on the dangers of unchecked technological power. And now it’s installing the very infrastructure those thinkers warned us about — on its own campus, against its own community.

And let’s talk about what these systems actually do in practice. Algorithmic classification by gender and age isn’t neutral. It’s wrong constantly. It misgenders. It misjudges. It encodes the biases of whoever labeled the training data into permanent, automated judgments about real people walking to class. When a human security guard makes a mistake, you can talk to them. When an algorithm tags you, you’re a data point in a system that doesn’t know it’s wrong and doesn’t care.

A camera that sorts you into categories doesn’t see you. It erases you and replaces you with a label. That’s not security. That’s automated diminishment.

The deeper problem isn’t MIT. The deeper problem is the normalization. Every institution that installs these systems makes the next installation easier. Every campus that accepts algorithmic surveillance makes it harder for the next campus to refuse. The overton window shifts, and what was once dystopian becomes standard, and what was once standard becomes mandatory.

Today it’s 500 cameras at MIT. Tomorrow it’s every university. Then every hospital. Then every public square. And by the time we ask whether this was a good idea, the infrastructure is permanent, the data is already collected, and the question is no longer whether we should be watched — but whether we even remember what it felt like not to be.

Every surveillance system is sold as a shield. But shields don’t have databases. Shields don’t classify. What MIT is building isn’t a shield — it’s a net. And the fish never vote for the net.

If you’re a student, a faculty member, a staff worker, an alum — this is your fight. Not because MIT is uniquely evil, but because MIT is uniquely positioned. If the institution that literally wrote the book on the dangers of algorithmic power can’t resist the urge to surveil its own people, what hope does anywhere else have?

Ask the hard questions. Who has access to the data? How long is it stored? What happens when law enforcement subpoenas it? What happens when ICE asks for it? What happens when a researcher wants to use your movement patterns for a study you never consented to? What happens when the system misidentifies someone and that someone is Black, or brown, or disabled, or just wearing the wrong color jacket on the wrong day?

The question was never ‘does this make us safer.’ The question is ‘safer for whom, and at what cost, and who gets to decide.’ If you weren’t asked, the answer is already clear: it’s not you.

500 cameras. $3 million. One institution betting that you’ll trade your privacy for the promise of protection. The bet is that you won’t notice, or won’t care, or will be too busy to object.

Prove them wrong.