My fieldwork has now brought me to Hyderabad, where I am conducting research with India’s most technologically advanced police force. I look forward to developing a deeper, more contextually grounded understanding of the social life of algorithmic evidence in criminal investigations and prosecutions

As part of my PhD fieldwork, I have joined Bureau of Police Research & Development - India’s state-owned policing think tank. I’ll be working as a research assistant for Tech & Modernisation Division - looking at use of AI in criminal investigation and evidence construction

Proposed changes to tax laws are simply providing a remote legal backing to an already prevalent police practice - confiscating devices and roving inspection of personal chats

Defence AI is built by, and for, war capitalists.

Authoritarian regimes have not just Co-opted “decolonisation” and “sovereignty”, but also “privacy” and “data protection”. State is claiming privacy against its own citizens.

Delhi Police arresting people based on facial recognition match score of 32%! At this point, such machine-driven ‘intelligence’ is simply a legitimation/rationalisation exercise.

Thinking of Karen Barad’s relational ontology. ‘Identification’ here is no longer a search for pre-existing reality but a discursive tool to actively shape and produce reality. The reality here being “criminality”.

While aggregated data on attacks on police is publicly available, the same for attacks by the police is nowhere to be found - and now being actively denied under “right to privacy” of the officers involved. Classic case of weaponising liberal reforms discourse

Theft in India is punishable with up to 2 years in jail, deemed a petty offence as per settled law. Arrest should only be made in exceptional circumstances. But FRT cameras has turned this upside down - amplifying the criminalisation and detention of those accused of petty, not serious offences

Indian state of UP spent 400 million INR to procure “AI crowd management tool” which was only used for counting footfall. Despite this, a tragic stampede took place where hundreds (still counting) lost their lives. Later, more money was spent in pushing positive ads for “digital kumbh”

Love archival research - as someone who does empirical research in criminal justice, archives (to an extent) are the least triggering, traumatising, and exhausting, source for data. Fun fact: colonial archives are more easily accessible than postcolonial ones 💀

Digital infrastructures in policing - both materially and symbolically - are mobilised arbitrarily. ‘Efficiency’, thereby, both in its functional and discursive operation, is multivalent, fluid, contingent to the context in which it is invoked - becoming complicit in punitive-political policing

From the Field: “We don’t use them much, given Delhi’s traffic, we prefer bikes or our normal cars … they’re too big, if we use them, we’ll never reach the spot in time” … “but they look cool, right? mostly for the show, I’d say” - patrol officers on the new “mobile” high-tech fleet.

Notes from the field: immune (so far) from race to algorithmic policing, this police station in a hilly state in India still uses pin-maps for patrol allocations

Notes from the field: I was surprised to see how the neatly drawn boundaries of public-private divide collapse in digital policing projects. (1/2)