Watching ourselves: Labour, Blair, and the Return of the Panopticon

For a century the ‘panopticon’ has stood as a symbol of oppression and paranoid surveillance. For a politician to be talking about it in this day and age underlines no a policy but a political mindset.

Politics often reveals itself not through legislation, but through the metaphors those in power reach for when they think they are speaking freely. When Home Secretary Shabana Mahmood suggested that artificial intelligence could finally realise Jeremy Bentham’s Panopticon – allowing ‘the eyes of the state to be on you at all times’ – she was not outlining a policy so much as revealing a political imagination.

It matters not only because of what is being suggested, but because of who is suggesting it, when, and in whose company. A Labour government, barely settled into office, openly invoking one of the most infamous symbols of surveillance and discipline in political theory – and doing so in conversation with Tony Blair – should give us pause. Even if this remains, for now, a thought experiment rather than a legislative blueprint, the fact that it is being mused about tells us something important about the political instincts shaping the current government. 

To understand why this matters, it is worth making a brief detour into theory – not as academic indulgence, but because Bentham and Foucault give us a vocabulary for recognising power when it presents itself as common sense. 

Bentham’s Panopticon was, on its face, an efficiency project. Designed in the late eighteenth century, it was a circular prison with a central watchtower, allowing a small number of guards to observe a large number of prisoners. The prisoners could never see the guards, only the tower. Crucially, they could never know when they were being watched. The result was not constant surveillance in practice, but a constant possibility of surveillance – enough to induce prisoners to regulate their own behaviour. 

Bentham believed this was humane. There was less need for brute force, fewer guards, and lower costs. Discipline became internal. The prisoner behaved as if they were always being watched, even when no one was there. In Bentham’s mind, this was progress. 

Michel Foucault, writing almost two centuries later, understood the Panopticon less as a prison design and more as a political technology. In Discipline and Punish, he argues that the true power of the Panopticon lies in its portability. Once you grasp the principle – visibility without reciprocity, observation without accountability – you can apply it anywhere: schools, hospitals, factories, offices, welfare systems, borders. The modern citizen becomes both prisoner and guard, monitoring themselves according to norms they did not choose and rarely question. 

This is where Mahmood’s remarks become genuinely alarming. The Panopticon is not controversial because it is old-fashioned or crude, or because it works too well. It replaces overt coercion with self-discipline. It makes power quieter, cheaper, and harder to resist. You stop because the light is red even when the road is empty. You comply because you might be observed. Authority no longer needs to shout. 

When a home secretary speaks approvingly of this logic in the context of AI and criminal justice, the question is not simply who will be surveilled – prisoners, parolees, suspects – but how far the logic travels once embedded in state infrastructure. Surveillance systems, once built, have a habit of expanding. They migrate. They normalise. They quietly reframe what is reasonable. 

This is why the familiar defence – ‘if you’ve done nothing wrong, you have nothing to hide’ – is so hollow. Privacy is not about secrecy; it is about power. 

In legal terms, this intuition is reflected in Article 8 of the European Convention on Human Rights, incorporated into UK law through the Human Rights Act. Article 8 guarantees the right to respect for private and family life, home and correspondence. Crucially, it treats privacy not as an individual preference, but as a structural limit on state power. 

While the right is qualified, any interference must be lawful, necessary, and proportionate. Panoptic surveillance cuts against those principles at their core. A system designed around constant or ambient visibility does not respond to specific suspicion; it normalises intrusion as a backward condition of social life. It blurs the line between being investigated and simply existing. 

Once surveillance becomes infrastructural rather than exceptional, it becomes almost impossible to meaningfully assess proportionality. The harm lies not only in what the state does with the data, but in how the expectation of observation reshapes behaviour long before any formal interference occurs. 

It exists to restrain the state, not to reward good behaviour. History shows us, again and again, that mass surveillance does not fall evenly. It concentrates its harms on marginalised communities, on political dissenters, on those already deemed suspicious.

And then there is Tony Blair.

It is difficult to separate Mahmood’s comments from the wider ecosystem of ideas in which they are circulating. Blair’s presence is not incidental. His record on civil liberties – from counter-terrorism powers to ID cards – speaks for itself. But what makes this moment distinct is the convergence of surveillance, AI evangelism, and private power. 

“before surveillance becomes law, it becomes thinkable. And before it becomes thinkable, it becomes imaginable – framed as sensible, modern and unavoidable.”

Through the Tony Blair Institute, Blair has become one of the most influential advocates of data-driven governance in the UK and beyond. In principle, using data better to improve public services is uncontroversial. In practice, the lines between public interest, political access, and corporate profit have become dangerously blurred. 

The Institute’s close relationship with Larry Ellison, founder of Oracle, raises uncomfortable questions about who stands to benefit from a centralised national data infrastructure. Britain’s NHS holds one of the most comprehensive population health datasets in the world. Its value – economic, political, strategic – is enormous. Calls to unify, centralise, and externalise this data are not neutral technical fixes. They are political choices about ownership, control, and accountability. 

From a Foucauldian perspective, this is panopticism updated for the twenty-first century. Data replaces walls. Algorithms replace guards. What changes is not merely the technology, but the location of discipline. Control no longer needs to be exercised through visible institutions of punishment. It is embedded into systems of access, eligibility, risk scoring, and prediction. Governance becomes anticipatory rather than responsive, acting on what people might do, rather than what they have done.  The watchtower becomes invisible, distributed across platforms, contracts and cloud services. The citizen is rendered legible, sortable, and predictable. Power no longer needs to ask; it already knows.

What is striking is how often this is framed as inevitability. AI is coming. Surveillance is coming. Growth depends on it. Productivity demands it. Resistance is cast as naive, nostalgic, or anti-progress. But inevitability is one of power’s oldest tricks. It discourages scrutiny. It short-circuits democratic debate.

This is where the symbolism of Labour’s flirtation with the Panopticon matters most. Labour was once the party that understood – instinctively – that state power must be constrained, not merely well-intentioned. Those rights exist precisely to limit what governments can do, even in pursuit of efficiency or safety. That trust, once eroded, is difficult to rebuild. 

Instead, what we are seeing feels like a familiar pattern: a party in government reaching for technological solutions to political problems, leaning on surveillance rather than solidarity, management rather than consent. The language may be softer than that of the right, but the architecture of control looks uncomfortably similar. 

None of this is to deny that technology can be used for good, or that the criminal justice system needs reform. The question is not whether AI can play a role, but under what conditions, with what safeguards, and in whose interests. A society organised around constant observation is not neutral. It changes how people behave, speak, organise, and dissent. It produces conformity not through fear of punishment, but through fear of visibility. 

Perhaps the most telling irony is that the state is increasingly relaxed about watching us, but deeply uncomfortable when we watch back.  This is where the concept of sousveillance matters. If surveillance is power exercised from above – the many being observed by the few – sousveillance is its inversion: citizens documenting, recording, and scrutinising those who govern them. It is one of the few mechanisms through which democratic accountability can be exercised in an age of asymmetric visibility. 

Yet sousveillance is routinely framed as dangerous, disorderly, or illegitimate. Filming police officers, tracing political donations, exposing lobbying networks or conflicts of interest are treated as threats to privacy or public order. The same institutions that insist constant observation is harmless when directed downward suddenly rediscover the sanctity of privacy when scrutiny moves upward. 

This asymmetry is not accidental. Panoptic power depends on one-way visibility. It functions best when citizens are legible to the state, while the state remains opaque to citizens. Sousveillance disrupts this arrangement by reversing the gaze – and it is precisely this reversal that makes it politically intolerable.

Foucault warned that we rarely suspect the conditional nature of our perception. The danger of the Panopticon is not that it forces us to behave, but that it trains us to believe this is simply how the world is. That being watched is normal. That questioning it is irrational. 

Labour does not yet have a Panopticon policy. But has something more revealing: a willingness to speak its language. That alone should concern us. Because before surveillance becomes law, it becomes thinkable. And before it becomes thinkable, it becomes imaginable – framed as sensible, modern and unavoidable. 

We should be resisting that framing, not after the infrastructure is built. The question is not whether the state can watch us more closely, but whether it should – and who gets to decide. If Labour cannot answer that convincingly, then the problem is not a single remark, but a deeper drift in how power is understood and exercised. 

The Panopticon was never just a prison design. It was a warning about what happens when efficiency is allowed to outrun accountability, and when power learns it can govern best by being unseen.