In solving crime, the false promise of technology

The central assumption underlying the Criminal Procedure (Identification) Bill, 2022 (“Bill”) is that more technology will help solve crimes more effectively and reliably. It permits forced collection, indefinite storage, profiling, databasing and sharing of a wide range of physical, biological and behavioural “measurements” for undefined purposes. The list of measurements in the Bill includes fingerprints, footprints, palmprints, photographs, iris and retina scans, signatures and handwriting. Additionally, the category of “biological samples and their analysis” could potentially include any bodily substance or feature, with any testing done on it, irrespective of its forensic value. Similarly, “behavioural attributes”, which remain undefined, could cover any kind of psychological, neuro-psychological or physiological examination connected to understanding a person’s behaviour. Given its scope, the debate on grave constitutional concerns must be informed by a close analysis of its technological premise.

Technology can go wrong. The Bill creates a system with virtually no fetters on taking and processing of measurements from any arrestee, detainee, convicted offender, or any person whose measurement may be considered “expedient” (i.e. convenient or useful) for any investigation. Before building such a vast system, one should consider if there are any scientifically valid techniques with recognised standards for analysing the wide range of measurements covered. The Bill appears oblivious to the growing body of scientific research which raises concerns regarding the validity and reliability of many pattern matching techniques for comparing measurements. Even the “advanced countries” that the Bill refers to are taking cognisance of these concerns to amend their own investigative and evidentiary practices. For instance, the 2016 report of the US President’s Council of Advisors on Science & Technology analysed existing scientific literature to examine the validity of six forensic methods, including fingerprint comparison, for which it noted the possibility of a false positive rate of one in 18 cases. Similarly, the American Association for the Advancement of Science in 2017 concluded that fingerprint comparison is prone to error and contextual bias, and there is no scientific proof to state that every individual’s fingerprints are unique. Following these developments, the US Department of Justice issued guidelines for latent print examiners, prohibiting them from making claims that two prints originated from the same source, or that latent print examination is “infallible with zero error rate”. Therefore, before embarking on a mindless expansion of collecting measurements, consider the existing scientific proof (or lack thereof) for the level of individualisation possible from using these measurements and whether it would support the “unique identification” of perpetrators.

The Bill justifies building vast databases of different types of measurements since they are “unique” to individuals and would, therefore, allow identification of the offender through comparison with evidence collected during investigations. The usefulness of a forensic database is guided by the level of individualisation possible by running the unknown samples against the entries of known individuals in the database. However, for several measurements covered within the Bill, such as footprints, palmprints, iris and retina scans, signature and handwriting, there is no data or scientific research to calculate their rarity and substantiate their discriminatory value. For example, if an unknown handwriting sample shows a million possible hits on the database, what would be the utility of such a search? While arguably, reliable comparison between the features of the suspect with the unknown sample collected during investigations may have corroborative value, this is already underway by collecting measurements from arrestees in individual cases. There exists no justification for building the massive forensic databases envisaged under the Bill.

The Bill also ignores the fact that forensic evidence can often be irrelevant or even misleading, depending on the context of the case. Instead of requiring probable cause to support the collection of measurements, the Bill grants unguided discretion to police or prison officers and magistrates, which may not assist reliable and fair investigations in any way. This gains significance as there are no established scientific protocols for forensic examination or clear legal standards for examination of expert evidence. Instead of directly addressing the issues with India’s weak forensics infrastructure, the Bill only intrudes upon privacy behind the ruse of promoting security and aiding investigations.

Shreya Rastogi heads the work on forensics and death penalty litigation at Project 39A, National Law University, Delhi The views expressed are personal

This article first appeared in The Hindustan Times and can be accessed here.