Asha Kapoor
When the state can identify every face, protest begins to change before anyone is arrested.
New Delhi, India
A protest has always required visibility. People gather because they want to be seen, counted and heard. Yet something changes when visibility is no longer directed only toward fellow citizens, journalists or political institutions, but toward machines capable of converting a face into an identity, an identity into a record and a record into a reason for further scrutiny.
That change is becoming tangible at Jantar Mantar, one of India’s most recognizable spaces for political dissent. Delhi Police have deployed live video feeds, mobile surveillance units and AI-assisted facial recognition during demonstrations, arguing that the technology helps identify people with criminal records and maintain public order. Protesters have responded differently, with some covering their faces not because they reject visibility, but because they no longer know what visibility may become.
This is where the politics of surveillance becomes more complicated than the familiar debate over privacy. A face at a protest is not merely biometric data because it belongs to someone occupying a specific social and political position. The same camera may record everyone equally while the consequences of being recognized remain profoundly unequal.
India’s constitutional tradition protects speech, assembly and privacy, but digital surveillance introduces a different form of pressure because it does not need to prohibit protest explicitly. The knowledge that participation can be recorded, matched and retained may be sufficient to alter behavior. A citizen can remain formally free to demonstrate while becoming less certain about what that decision will mean months or years later.
That uncertainty is difficult to measure because it rarely appears in arrest statistics or produces an immediate constitutional crisis. It appears instead in smaller decisions: whether to attend, whether to cover a face, whether to remain after police cameras arrive or whether participation today could create consequences in another institutional setting tomorrow.
For women, caste-marginalized communities and religious minorities, that calculation can become more complex. Surveillance technologies enter societies that already contain unequal forms of exposure, authority and vulnerability. An algorithm does not need to understand caste, gender or religion explicitly to operate inside institutions where those categories already shape encounters with police, bureaucracy and public judgment.
This is why the argument that facial recognition merely identifies known offenders remains incomplete. Delhi Police told the Supreme Court that their system identified thousands of people with past criminal records during earlier protests and rejected allegations of indiscriminate profiling. Even if those claims are accepted, a broader democratic issue remains: what happens to public assembly when thousands of faces can be compared against state databases in real time?
Modern cities require investigative capacity, and political demonstrations can create genuine security challenges. The harder issue concerns what happens to dissent when identification becomes ambient rather than exceptional, operating continuously in the background of civic life. A technology can remain formally lawful while still changing the behavior of people who know they are being watched.
India may therefore be entering a period in which participation and traceability become increasingly difficult to separate. That does not automatically amount to repression, just as a camera does not automatically become an instrument of authoritarianism. The transformation may be quieter, emerging through the accumulation of choices made by citizens who begin calculating the future consequences of being visible before deciding whether to speak.
The implications extend beyond one protest site or one police system. Once biometric identification becomes embedded in public space, the distinction between security infrastructure and political infrastructure becomes harder to maintain. The same system that recognizes a wanted suspect can also recognize a dissenter, a student or a citizen whose only meaningful act was deciding to stand in a crowd.
Democracy has always depended on people entering public space without knowing whether they will prevail. It becomes something different when political participation also requires weighing how long the state may remember a face, where that recognition may travel and what future meaning might be attached to a moment of dissent.
Asha Kapoor | South Asia Strategic Affairs Columnist, Phoenix24 | Gender, Surveillance & Narrative Warfare