The Return Of The Criminal Face: Algorithmic Policing And The Revival Of Discredited Criminology

Nitya Jain
DOI:

Abstract

The facial-recognition technology, tightly integrated into modern policing procedures, has the appearance of being a dispassionate and objective enhancement of the investigative potential. Yet, at the heart of its veneer of computation, there is an underlying epistemic continuity with the discredited tradition of appearance-based criminology. In history, the effort to tell criminality by its physical appearance and most famously with Lombrosian physiognomy, collapsed due to its methodological weaknesses and inherent biases. However, in an altered form of this, the same inferential logics are reborn as algorithmic systems that translate facial features into probabilistic identifiers and suspects. The implementation of facial recognition in policing in India illustrates how the output of algorithms, despite being framed as statistical correlations, is realised as categorical truths. This is not only a technical change, but an institutional one- the probabilistic matches are promoted to actionable suspicion, usually without questioning the error rates or bias of the dataset and the validity of the methodology. By doing this, such systems pose the danger of formalising historical trends of over-police and social marginalisation into the very fabric of law enforcement. Placed in the shifting regulatory context of the Bharatiya Nagarik Suraksha Sanhita, 2023 and the Bharatiya Sakshya Adhiniyam, 2023, the article reveals a significant regulatory loophole- although both laws enable the admission and growth of digital evidence, they are conspicuously silent on the standards of algorithmic reliability, transparency and accountability. However, eventually, the uncritical embrace of algorithmic policing can become a threat to the institutionalisation of a new kind of technologically mediated prejudice. A paradigm shift is required, shifting from uncritical adoption to principled regulation and making it clear that the legitimacy of criminal justice is not to be found in automation but in its adherence to constitutional values of fairness, accountability and justice.

Keywords

Algorithmic Bias, Facial Recognition Technology, Digital Evidence, Physiognomy, Procedural Fairness, Constitutional Accountability