Course Correction on Delays and Bail under the UAPA?

[This post was updated at 6PM on 19.05.2026 to add reference to this news article]

On 18.05.2026, a Division Bench of the Supreme Court handed down its verdict in Syed Iftikhar Andrabi [2026 INSC 503]. The bail application was allowed, which if we see the facts is not entirely surprising: the accused had been in jail for more than five years, there were many co-accused already on bail, some of them quite similarly placed, the incriminating material was largely confessional, and the trial was nowhere near completion. 

It is not because of the factual analysis that the judgment runs into 102 pages, which is quite unusual for a bail petition. The length was necessitated as the Bench felt it had to clear the legal position on a key issue affecting bail adjudication in the UAPA context, and the broader context of special crimes more generally. This, of course, was the issue of what impact does lengthy undertrial incarceration have in bail petitions under the UAPA with its additional restrictions for grant of bail under Section 43-D.

According to the Bench, a position had been set out by the Bench of Three Justices in K.A. Najeeb [(2021) 3 SCC 713], holding that the rigours of merits-based statutory restrictions for granting bail would, in the case of grave delays, be trumped by the concerns for an accused's liberty. The value of liberty and to a speedy trial was seen as grounded in Article 21 of the Constitution; lengthy undertrial detention would, in case of grave delay, be seen as causing detention in a manner contrary to these fundamental rights.

This position seemed to have been doubted and departed from in two subsequent judgments delivered by Division Benches; the first was Gurwinder Singh [(2024) 5 SCC 403], and the second Gulfisha Fatima [2026 SCC OnLine SC 10] (both were opinions of the same author). These subsequent decisions had, seemingly, held that bail was not the rule in UAPA cases owing to the restriction placed by Section 43-D UAPA which evidenced a legislative preference for jail as the default, and no relaxation or mollification of this statutory restriction was caused by delays or by the lengthy undertrial detention of an accused. 

The validity of this line of reasoning from both these opinions has been expressly doubted in Syed Iftikhar Andrabi [Paras 26 to 41]. The simplest line of critique is that the Division Benches could not have doubted or disagreed with the judgment of Three Justices in Najeeb. But the Court does not rest there, perhaps due to the manner in which both Gurwinder Singh and Gulfisha Fatima made departures. Neither said that the law laid down in Najeeb was bad; instead, the fight was about what law, if any, did Najeeb lay down in the first place. 

Gurwinder Singh and Gulfisha Fatima departed from the opinion in Najeeb by reading the latter as tightly wound to its facts. In that case there was no prospect of speedy trial whereas in Gurwinder Singh notable progress had already been made, and in Gulfisha Fatima the delays were 'natural' owing to the complexity of the case. In other words, both opinions tried to denude the opinion in Najeeb of having laid down any law at all. Having done so, both opinions offered their reading of the law, which saw the statutory bar of Section 43-D UAPA remaining unaffected by any length of delay. Even if there was delay, the detention of an accused may yet be justified in the overall facts of a case.

Syed Iftikhar Andrabi has specifically disapproved of this interpretation of Najeeb, as well as the holding of both Gurwinder Singh and Gulfisha Fatima that the statutory bar is adamantine and remains unaffected by delays. It reminds us of the very clear observation in Najeeb, that in appropriate cases the statutory bar would 'melt' in the face of the pressure created by the prolonged and unconstitutional deprivation of an accused's liberty. The statutory bar is, and remains, subservient to constitutional interests.

The verdict is a helpful course correction by the Supreme Court in the domain of bail on grounds of delay which has seen a renaissance of sorts in the past five years across multiple statutory contexts containing a clause restricting grant of bail. Whether the course correction will hold or not is something to be seen. All it takes is another Bench of Three Justices to come and doubt Najeeb, after all. And it took only a day for the State to raise the possibility. Think about this for a moment. It is not a right to release on bail, but the mere possibility for considering bail pleas slightly more favourably for those who have spent inordinately long periods in custody pending trial, is what the fight is about. And even this possibility is precarious, dependent upon how some words in a statute are read. For now though, the possibility shines brighter than it did six months ago.

At the same time, let us not overstate the usefulness of this jurisprudence of apology created by our courts. Where, after depriving individuals of their liberty for years, at some unspecified point in the future their custodial detention mysteriously burnishes their claims to liberty with greater light. In fact, Syed Iftikhar Andrabi has ironically reinforced the sheer unfairness of it all quite drastically. At paragraph 42, the Court cites data suggesting extremely low conviction rates for UAPA crimes, and rhetorically asks how can the prolonged detention be justified merely because of serious allegations if there are such high possibilities of a future acquittal. If this reasoning is good for prolonged detention and delay on grounds of bail, it is also good for demanding greater accountability for deciding bail applications on merits and seriously assailing the constitutionality of the restrictive bail clauses under the UAPA and other laws.

The allegations against the accused were serious in Syed Iftikhar Andrabi, but they were based on hardly any credible material. So why did the courts below continue to shy away from granting bail on merits for the five years that it took for this judgment to pass? That, really, is the question which deserves a lot more scrutiny going forward in bail jurisprudence under the UAPA. 

The vigour with which the Court has sought to resolve the creases within the legal fabric on delay-based bail for UAPA and other such crimes requires urgent reciprocation in the realm of merits-based bail. Only then will the proclamation of the Court in Syed Iftikhar Andrabi carry any real force: that bail remains the rule, and jail the exception, in all circumstances in our law.

[On reading of restrictive bail clauses, readers may find the six-part series 'Jail is Not the Rule' helpful]

Snippet - The Prison Statistics India 2024 Report

This past week, the Prison Statistics India Report for 2024 ['PSI Report'] was published by the National Crime Records Bureau ['NCRB]. I could not spot any massively interesting data points as such, especially if we read this Report together with the 2023 PSI Report. Indian prisons continue to be filled overcapacity, mostly with undertrial prisoners, who are usually young and poorly educated. While we do know that the prisons were collecting both religion and caste demographics, these indicators continue to be out of sight in these Reports. Economic profiles of prisoners are also not provided. If we did have that data provided in the Reports, one suspects it would further confirm that our prisoners are, disproportionately, coming from marginalised backgrounds.

The five-year charts in the 2024 Report show us that there has been a tapering of the prisoner population. The highs seen in the 2022 Report have, thankfully, not been witnessed again. Any positivity generated by this data is offset when we consider that that reasons for that tapering is not systemic improvement. One big reason are the efforts made in 2023 by the Supreme Court to push for release of prisoners through undertrial review committees. This is coupled by consistent decrease in prisoner population seen in Uttar Pradesh over the past couple of years. Anybody can see that these are not long-term solutions.

The 2024 Report is going to be interesting historically, as hereon the legal framework will no longer speak to the Criminal Procedure Code of 1973 but the Bharatiya Nagarik Suraksha Sanhita of 2023. 1973 saw the introduction of a new Code with an impetus on ensuring speedy justice and provisions meant to reduce unjust undertrial incarceration. It went out in 2024 with an unpleasant legacy on these fronts. The new law has built upon the same rotten foundations but with some tweaks on aspects of arrest and bail, so it would be interesting to see what changes it brings on the ground. 

A case in point will be the impact on default bail under what was Section 436-A of the 1973 Code, which has now become Section 479 of the 2023 Sanhita. This scheme of default bail under the old Code had enabled release of prisoners who had served half their prison term. The revised scheme allows for release on serving one-third of the sentence, provided it is a first-time offender (with some more caveats which we needn't elaborate here). 

Implementation of 436-A is seen by some as a chokepoint resulting in unjust incarceration. However, the PSI data has consistently shown that this may not be the case, as there was hardly a sizeable number of the prison population eligible for release in the first place. The 2024 Report reaffirms this trend. As we have seen, 479 BNSS will change the eligibility criteria, potentially expanding the reach of the clause. It will be fascinating to see the impact it has. If the half-year's worth of data is anything to go by, the future of this species of default bail may not be very different from its past.

 

Thinking about Section 27 (and 8) of the Evidence Act

Come July 2026, it would be two years since the three new criminal codes were brought into force. That is hardly long enough for the heavy, centuries-long, shadows of the old codes to disappear from the scene. At some points, you even feel sad for the new laws, which nobody can get around to either understanding or applying without referring to that familiar array of old clauses and provisions. Take Section 23 of the new Bharatiya Sakshya Adhiniyam 2023 ['BSA'] for example. Will Section 23 ever have an identity or an existence independent of its illustrious ancestors, Sections 25, 26, and 27 of the Indian Evidence Act 1872 ['IEA']? I highly doubt it.    

Ruminating about the identity crisis that will plague the new criminal codes is not the point of this post, as fun a topic as that may be. My focus is on the proviso to Section 23 of the BSA, or more specifically, its predecessor in Section 27 of the IEA. It is prompted by a sobering realisation, that there has been no post on this topic so far on this Blog.

The Statutory Scheme

Sections 24 to 26 IEA placed different kinds of barriers on evidence of confessions. Section 24 rendered any confession given due to inducement or threat from a person in authority 'irrelevant'. Section 25 stated that no confession made to a police officer by a person accused of an offence 'shall be proved' as against him. Section 26 stated that no confession made while in police custody shall be proved against a person unless it was made in the 'immediate presence' of a magistrate. Then came the proverbial backdoor in the form of Section 27, providing that where any fact is discovered pursuant to information given by a person in the custody of police, so much of the information which relates distinctly to the fact discovered. may be proved. Note that this scheme has been condensed into two provisions under the BSA; Sections 22 and 23. Section 22 is the successor to Section 24 IEA, and Section 23 of the BSA contains the erstwhile Sections 25 to 27 IEA.   

Logic and Theory

It can be argued that there are two dominant rationales guiding the legal scheme on confessions: first, respect for the 'best evidence' rule, and second, a distrust of acts done or statements made by a person who is under police control.

The best evidence rule is a hallmark of modern evidence law theory, which has two components. Most of us are familiar with its court-facing part: courts ought to decide cases based on the best evidence available to them. There is a legislative corollary, too, which commands the legislature to tear down technical rules which stand in the way of courts accepting the best evidence. For its time, the Indian Evidence Act made a lot of progress on the legislative side of things by removing older common-law hearsay objections. Today, though, the scheme of the Indian Evidence Act again appears antiquated when tested on this anvil, as can be seen with the approaches taken in other jurisdictions over the 20th Century. And, with blind repetition of this antiquated 1872 scheme in 2023, the process of law on the whole continues to suffer (yet another digression). 

In India's context, the pursuit of best evidence for courts under the adopted English system has coexisted with a general distrust of evidence purportedly given by persons under control of the police. A legal bar on using any statements, including of ordinary witnesses, recorded by police during investigation came in place since the start of codification in the 1860s. Such a step, no doubt, reduced the evidence potentially available before a court to decide a case; but the decision to remove such material from the adjudicative exercise was consciously made owing to fears about the police creating the evidence in question, which rendered it highly unreliable. Such fears were at their highest when it came to evidence of confessions taken by police from the mouth of an accused, prompting the total bar on evidence of confessions made to police.

I have offered a very generic sketch of a fascinating historical development in the mid 19th Century, but it serves our purpose for now, which is to highlight the two rationales in play guiding the legal scheme on the evidence of confessions. Understanding these two rationales, we can now proceed to appreciate the compromise arrived at by Section 27 IEA adopting the logic of confirmation by subsequent events. On its own, a statement by an accused may be worthless; but what if it leads to discovery of facts? Excluding the evidence of the statement even where it leads to discovery was seen as a step too far. As a result, Section 27 IEA allows proof of only so much of the information given by an accused which relates distinctly to the fact discovered.  

Interpretive Questions

One may be forgiven to think that the existence of strict rules barring admissibility of confessions, except in limited cases, would have been enough for the Indian criminal process to gradually move away from trying to use an accused as the primary source of gathering incriminatory material. The shift never really took place, apart from one of nomenclature where confessions are now increasingly called disclosures. Even today, a classic criminal investigation will almost always involve the accused making a disclosure, which may or may not directly lead to any recovery. The continued reliance on confessions for gathering evidence by police has naturally meant that courts are repeatedly faced with questions about admissibility of such statements. Which, in turn, means interpreting Section 27 IEA and figuring out of any part of the confession fits its contours.

I want to focus here on three interpretive questions which courts have faced and answered over time: (i) the meaning of 'fact discovered' for the clause, (ii) scope of the 'information' that is rendered admissible, and (iii) relevance of the conduct of an accused in such scenarios. There is a surprising level of coherence in judicial interpretation around these aspects, with of course existence of the odd incorrect view taken by courts from time to time.  

Fact Discovered

Since the logical premise underlying Section 27 IEA is that the discovery of a fact ensures reliability of the information given by an accused in custody of police, it is perhaps unsurprising to note that courts are strict about reading 'discovery'. If the police already knew about something, then there is no discovery of that fact made pursuant to the information given by an accused. Where this causal link is snapped, courts will not invoke Section 27 IEA at all.

This is all fairly straightforward, until you travel into the realm of determining what is a 'fact' that stands to be discovered pursuant to a statement given by an accused. A person's state of mind is also a fact, so can we argue that this is a fact discovered from the information given by an accused? The answer is no. It is not difficult to reason why this argument does not, and should not succeed. In this case, the information by an accused is the fact discovered — there is no independent confirmation of the reliability of information by any subsequent event. To apply Section 27 IEA this way would negate the very purpose of the clause and allow admission of confessions wholesale. Thankfully, this is one of the areas where the life of the law has remained closely tethered to logic, and courts have consistently insisted that the fact discovered must be a tangible object having an existence in the real world independent of the information by an accused. A great place to go and read about the development of law in this context remains the judgment in Navjot Sandhu [(2005) 11 SCC 600] 

'So Much' of the Information     

A common mistake is to think that discovery of a fact renders the evidence of a confession admissible in its entirety. The clause is much narrower in scope, admitting only evidence of so much of the information by an accused which distinctly relates to the discovery, and nothing more. Faithful application of Section 27 IEA thus requires to things in turn: first, clear and cogent material to show what was the information given by an accused, as the Supreme Court recently reminded police agencies in Boby v. State of Kerala [2023 INSC 13]. Second, courts to run a scalpel over such information to excise everything other than so much of the information which relates distinctly to the fact discovered. A good example of this approach of wielding a scalpel in action is Mohd. Inayatullah [(1976) 1 SCC 828]. 

What about Conduct?

It is quite often that an accused person does not merely give a statement, but either points out a spot from where an object is found or, even better, actually produces it himself. At the same time, an accused may straightaway point to a spot or produce an object without making any clear statement about the facts. Such scenarios throw up a different kind of evidence for a court to consider: the evidence of conduct, which is separately relevant under Section 8 of the IEA. How to sort such situations out?  

Conduct mixed with statements is a problem catered to by the statute. Section 8 itself clarifies that conduct does not include statements, except those which explain conduct. So it cannot be that by using conduct we can somehow broaden the scope of what is permitted through the narrow confines of Section 27 IEA, and say that the statements themselves are independently made relevant. There is another reason for this; any statements to police are unusable in evidence by virtue of Section 162 of the Criminal Procedure Code 1973. Section 27 is a statutory exception to this bar, but Section 8 is not.

When it comes to conduct which is not accompanied by any statements, we are faced with the same issue of the best evidence rule tussling with the distrust of evidence gathered from an accused by the police. Intuitively, an accused leading the police to material objects connected to the case is a relevant fact, which law should allow evidence for. But the risks of police creating evidence, which we saw associated with the evidence of statements made to police officers, do not magically disappear if we replace statements with conduct. 

Courts have held that these tests remain as important where police claim that the accused showed us a spot, or led to a recovery, although there is no comparable statutory prohibition in this scenario. As a rule, we see courts insist on some discovery of fact to render evidence of conduct relevant and admissible, and analyse whether it was truly the conduct which led to the discovery or was it a fact already known to the police. Additionally, courts also repeatedly insist that conduct evidence alone is quite a poor indicator for proof, and requires solid corroboration for it to further along the chain of circumstances being created in any case.                  

Constitutional Conundrums?

After the introduction of the Constitution in 1950, the evidentiary scheme of confessions has come within the constitutional crosshairs on a two notable occasions before the Supreme Court (there have been some High Court decisions as well). 

The first of these was Deoman Upadhyay [AIR 1960 SC 1125] where the validity of Section 27 was questioned on grounds of arbitrariness. The clause only applies to an accused person in police custody, prompting the unsuccessful argument that this invidiously discriminates between those in the custody of police and others. The second, was Kathi Kalu Oghad [AIR 1961 SC 1808]. Unlike Deoman Upadhyay, the constitutional question was only indirectly raised, as the Supreme Court was asked to decide whether there was a presumption of coercion accompanying statements made in police custody, thereby rendering Section 27 IEA contrary to the right against compelled self-incrimination. The Court unanimously rejected this argument and insisted that mere interactions with the police could not be deemed coercive [Note, though, that 15 years later a different albeit smaller bench in Nandini Satpathy did consider such issues more favourably].

The Future

It is quite likely that the current state of affairs will remain undisturbed in the short-term future; for no reason other than the legislative call to stick with the antiquated IEA approach even in the 21st Century. It will be interesting, though, to see what happens in the next 10-15 years as more interactions with the police have the benefit of audio-video recordings. Will a contemporaneous recording of the interaction prove to be the antidote for centuries of distrust regarding police affairs? With sufficient safeguards in place, there is no reason for it not to. Once this hurdle is crossed, the underlying basis of Section 27 will also be open to contestation. But not before that.

Snippet: Bellin's Critique of Criminal Procedure doctrine as a Law of Cooperation

I came across a new article posted to SSRN by Prof. Jeffrey Bellin titled 'Criminal Procedure as the Law of Cooperation' (forthcoming in the Georgia Law Review, 2026). The paper is focused on the development of U.S. criminal procedure by their Supreme Court, and so is not directly dealing with things that this blog talks about usually. But I thought that the argument made in the paper resonated quite well with critiques of how the Indian judiciary has developed criminal procedure as well.

His claim is straightforward - it is maybe a mistake to think that the main path taken by American courts on developing criminal procedure law is a rights-affirming one, which imagines restrictive interpretations as anomalies. He argues that, if anything, the dominant path has been imagining criminal procedure as a law for securing cooperation with law enforcement. Cooperation is the primary goal, and the enforcement of rights is subservient to ensuring that this primary goal is not unduly compromised. Decisions actively affirming rights while truncating cooperation-based interests are few and are outliers; these have often been whittled down over time because of the dominant cooperation-based framework (Miranda being the example he offers). 

This critique connects with existing critiques of Indian Supreme Court jurisprudence. For instance, Mrinal Satish and Aparna Chandra had made a similar claim while surveying Supreme Court decisions on the fundamental rights linked most closely to criminal procedure in their 2016 essay 'Criminal Law and the Constitution' which was part of the Oxford Handbook. 

In the context of Indian law, one feature which adds to the complexity of the argument, is that unlike the U.S. where criminal procedure is largely a creation of judicial interpretation of the relevant clauses in the Bill of Rights, in India the law on criminal procedure remains rooted in statute, and this statute influences how the courts develop doctrine. A key focus of statutes governing criminal procedure would naturally be on maximising cooperation with agencies. This places it in conflict, somewhat, with other laws seeking to secure individual rights even at the cost of cooperation. Whenever there is a conflict, it is no surprise that Indian courts tend to read the law in a way which ensures maximal cooperation, with some outliers over the years. 

Those interested may find the piece engaging and as a useful resource to turn to for developing arguments about the doctrinal development of Indian criminal procedure.

Guest Post: Criminal Status as a Constitutional Filter - What the Uttarakhand High Court's Order in Deepak Kumar Gets Wrong

(This is a guest post by Ammar Shahid and Syed Raiyyan)

The order passed by the Uttarakhand High Court (“The Court”) on 19th March 2026 in Deepak Kumar @ Akki v. State of Uttarakhand is a tight disposal of a petition that was asking for considerably more than it deserved. Three of four reliefs were turned down, and rightly so. However, the reasoning deployed against a prayer for police protection to Deepak Kumar is receiving significant criticism. It is not that the Court reached the wrong result. It is the path it chose to get there that rests on a doctrinal footing that does not hold up under scrutiny. When a Court ties the availability of relief under Article 21 to a Petitioner’s status as a “suspected accused,” it makes a statement about constitutional entitlement that no precedent supports.

Factual Backdrop

On Republic Day, 2026, Deepak Kumar (“Petitioner”), a gym owner from Kotdwar, Uttarakhand, intervened when a group entered the shop of one Vakeel Ahmed. He was a seventy-one-year-old shopkeeper and was pressured into removing the word “Baba” from the shop’s name. When the crowd demanded to know who he was, Petitioner identified himself as “Mohammad Deepak.” Days later, a First Information Report (“FIR”) was registered against him and his associate on charges including misconduct, snatching a mobile phone, and criminal intimidation. A mob later gathered near his gym and shouted slogans against him. Petitioner filed his own complaints, and his gym membership dropped from approximately 150 to nearly 15 in the weeks that followed. It was in this backdrop that he approached the High Court under Article 226.

The Petition and its Reliefs

The Petitioners were accused persons in FIR No. 19 of 2026, registered under sections 115(2), 191(1), 351(2) and 352 of the Bhartiya Nyaya Sanhita, 2023. They approached the Court with four prayers: quashing of the impugned FIR; directions to register FIRs on their own complaint against alleged perpetrators of hate speech; police protection for themselves and their families; and a departmental inquiry against certain police officials for alleged partisan conduct.

The Court dismissed reliefs (ii), (iii), and (iv) at this stage. Relief (i), the quashing of the FIR, was listed for further arguments. On reliefs (ii) and (iv), the reasoning is difficult to fault. On relief (ii), the petitioner had suppressed the registration of two FIRs on his own complaint. Both were registered at the same police station as the impugned FIR, and both predated the writ petition by more than a fortnight. Additionally, the statutory remedy under Section 175(3) of Bharatiya Nagarik Suraksha Sanhita, 2023 (“BNSS”), 2023 was available and unused. On relief (iv), a prayer for departmental proceedings against investigating officers while one’s own investigation is still pending practically invites the inference of an attempt to complicate the inquiry. The Court’s scepticism there is entirely well-founded and supported by reasoning. It is relief (iii), the prayer for police protection, that demands a harder look.

The Court’s Reasoning on Relief (iii)

The Court refused police protection primarily on the ground that a “suspected accused” who is under investigation cannot pray for adequate police protection, and that the police administration is “competent enough to keep on watch for his safety.” The Court also noted that the concerned officer had conveyed by telephone that there was no credible threat perception. The last ground of there being no credible threat was a perfectly adequate basis to refuse the relief. The first ground, on the other hand, is constitutionally tenuous.

Entitlement to Relief vs. Merits of Relief

There is a distinction in adjudication that has been overlooked in the order: the distinction between a petitioner’s entitlement to seek a particular relief and whether the petitioner has made out a case for that relief on the merits. These are entirely separate questions, and treating them as one produces reasoning that goes further than any given case requires.

Entitlement asks: Is this person, as a matter of constitutional or legal principle, capable of invoking this relief at all? Merits ask: has this person placed sufficient material before the Court to justify granting it? A Court can answer “yes” to the first and “no” to the second, and proceed to deny the relief without any doctrinal baggage. What a Court cannot do is answer “no” to the first question when the relief in question flows from Part III of the Constitution. Article 21 has long been understood to encompass an affirmative obligation on the State to protect it. In Akansha v. State of U.P., the Allahabad High Court held precisely this. The right to protection of life is a constitutional obligation of the State, and this obligation runs independently of the petitioner’s conduct or legal status.

Consider how this plays out in practice. When the Court says that a “suspected accused” cannot pray for police protection, it is not just deciding this case. It is articulating a principle. That principle, taken at face value, would mean that any person against whom an FIR is registered loses the ability to approach a Court for protection of their life until the investigation concludes. The logical endpoint of that proposition is uncomfortable: a person facing genuine, credible, life-threatening danger from a third party would be turned away at the door of the writ Court simply because he is simultaneously accused in another matter. Yet it is what the reasoning in this order, read literally, imports.

The merits dimension, by contrast, was entirely workable on the facts of this case. The petitioner placed no concrete material before the Court demonstrating a real threat. The investigating officer himself reported no credible threat perception. In those circumstances, the Court had everything it needed to deny the relief on the merits: there was simply nothing on the record to justify invoking the Court’s extraordinary writ jurisdiction for police deployment. The entitlement question did not need to be answered, and answering it the way the Court did created a proposition that the facts of the case did not require. This distinction is not novel, and the Courts have been asked to apply it in precisely this context before.

What Constitutional Precedent Actually Establishes

The Supreme Court’s Article 21 jurisprudence has maintained one consistent thread: the right to life does not shrink based on a person’s status in criminal proceedings. Several cases make this unavoidably clear.

In Arun Gavli v. State of Maharashtra, the Bombay High Court addressed this question of entitlement in the context of a person with a formidable criminal history. The Court’s approach was that the State’s obligation to provide protection under Article 21 is not neutralised by the fact that the person threatened carries FIRs or pending cases against him. The inquiry must remain limited to the threat, not the character of the person threatened. The Court held that “if a person like the petitioner, makes out a specific case of a genuine and imminent threat to his life on a particular occasion or at a particular place, the State Government would not be justified in refusing to grant armed police protection to him qua that particular occasion or place. This would, however, be subject to occasional review by the Threat Perception Committee constituted by the State Government.”

The Madras High Court’s position in K. Kalaivanan v. Superintendent of Police, Thiruvarur District is to a similar effect. The petitioner was involved in criminal cases and sought protection from rival group threats. The Court, in directing appropriate protection, reiterated that the right to life does not stand diminished by the existence of criminal proceedings against a person. Further, it is for the police to assess the credibility of the threat on the ground and not for the Court to bar the gateway to relief based on the petitioner’s legal status.

In Sunil Batra v. Delhi Administration, the Supreme Court held that even a convict retains fundamental rights. The Court drew a clear line: lawful custody restricts only those rights that cannot be exercised because of confinement itself. All other rights remain intact. If a convicted prisoner retains rights under Article 21, the argument that a “suspected accused,” who benefits from the full constitutional presumption of innocence, has a diminished entitlement to seek protection of his life would be legally untenable.

Taken together, these decisions convey what the Uttarakhand High Court’s order does not engage with. Courts have been asked this question of entitlement several times and have answered it in the affirmative. The identity, criminal status, or legal record of the person seeking protection goes, at most, to the credibility of the threat claim. It does not operate as a threshold disqualification.

Conclusion

To be fair to the Court, the outcome on relief (iii) is defensible. The petitioner placed no material before the Court establishing a genuine, immediate, or concrete danger. The police’s own assessment was that no credible threat existed. The relief simply was not made out on the facts. These were sufficient, clean, and principled grounds to deny the prayer.

However, the path the Court took to get there is a path that constitutional precedent does not sanction. Precedents converge on the same point: criminal status is not a constitutional filter through which the right to life must pass.

Courts often deal with petitions that are, as this one appears to have been, partly instrumental. The instinct to reject them firmly is sound. However, the legally sound path is in rejecting them on the ground that corresponds to the actual problem of insufficient merits. Formulating a broader principle that could one day be applied against someone whose threat is entirely real and whose only misfortune is the existence of an FIR bearing his name is untenable. One ground was enough here. Two grounds, when one is constitutionally unsound, is one ground too many.

Guest Post: On the Allahabad HC Direction for Mandatory Prosecution of Complainants in Closure Report Cases

(This is a guest post by Vaneet Kumar)

The criminal process in cognizable offences is ordinarily kick-started by filing a First Information Report (“FIR”) under Section 173 of the Bharatiya Nagarik Suraksha Sanhita (“BNSS”). This is followed by the police investigation, which culminates in a Police Report where police recommends either launching a prosecution or closing the case. The report in the former scenario is commonly referred to as a Chargesheet or Challan across many parts of India; and in the latter scenario, it is called a Closure Report. 

On 14 January 2026, the Allahabad High Court in Umme Farva v. State of Uttar Pradesh ordered that, as per law, where the police file a closure report they must also launch a prosecution against the complainant by way of a written complaint under sections 212 and 217 of the Bharatiya Nyaya Sanhita of 2023 (“BNS”). Subsequently, this decision was appealed before the Supreme Court, and the Court in Mahmood Ali Khan v. State of Uttar Pradesh has put this direction on hold, for now. A good time as any to discuss the issue.

Examining the Sources of the HC Direction

Sources/Reasoning Behind the Direction

Umme Farva frames this direction to file a mandatory complaint as a “statutory obligation” of the investigation officer, the failure to dispose of which will not only attract an offence under Section 199(b) of BNS (para 15) but also departmental proceedings and contempt case (para 21). The High Court framed this stringent and deterring obligation on the collective reading of the provisions of substantive and procedural criminal laws dealing with the offences of false complaint. It began by emphasising that Section 215(1) of the BNSS forbids a court from proceeding in matters under Sections 212 and 217 of the BNS without a “complaint in writing of the public servant concerned” or its superior (para 16). Therefore, it reasoned that retaining the filing of a complaint as discretionary would render the purpose of Sections 212 and 217 of the BNS redundant and defeat the legislative intent (para 47). After passing this general direction for the IO, proceeds to integrate it into the judicial system. It reiterates the established position of law: a magistrate, upon receiving the Closure Report, can still take cognisance, order further investigation under section 193(9) of BNSS, or consider a protest petition by the informant (para 31). As a result, the mandatory police complaint under sections 212 and 217 of the BNS “shall be kept in abeyance till the acceptance of the final/closure report or rejection of the protest petition/application” (para 34). Thus, only after the magistrate believes that no offence can be made out against the accused, the magistrate should go ahead under sections 212 and 217 of the BNS (para 31).

Analysis and Application of Existing Jurisprudence

There is not a lot of law interpreting the BNS, so we would need to look at the position as it availed for these offences under the repealed Indian Penal Code of 1860 (“IPC”). The predecessors of Sections 212 and 217 of the BNS were Sections 177 and 182 IPC, respectively. Analysing the precedents on these provisions, it would seem that the directions in Umme Farva are directly contrary to the Supreme Court’s observations on the matter. The Supreme Court in Daulat Ram v. State of Punjab noted that these offences do not require that action must be taken. It recognised that there is a discretion, rather than an obligation, vested with the police to initiate action on a false complaint. Similarly, Basir-Ul-Huq v. State of West Bengal also contemplates and permits the scenario where the police does not take any action under these sections. 

Besides being apparently contrary to judicial precedent, I would also argue that the discretion-based approach which courts had hitherto supported also makes a better argument in principle. The High Court in Umme Farva reasoned that Section 199(b) BNSS can't be seen as giving any discretion, as it would in turn nullify the offence itself. But the very idea behind requiring a formal complaint from a specified officer was precisely to prevent every case of an acquittal or closure from becoming a prosecution against the complainant. In State of Uttar Pradesh v. Mata Bhikh, the Supreme Court held that the object of the requirement for a complaint “is to protect persons from being vexatiously prosecuted upon inadequate materials or insufficient grounds”. Thus, the requirement of complaint is not to initiate automatic prosecution to avoid nullification of BNS provisions, but to protect the complainant from frivolous proceedings. Unnecessary prosecution in bona fide complaints that the bar originally intended to keep at bay will amount to mala fide prosecution, even though it is a product of judicial diktat. Moreover, the unclear jurisprudence on the state’s responsibility for victim compensation in cases of mala fide prosecutions adds to the worries of the original complainant (page 44). And it is not as if accused persons are remediless. An accused person, in the absence of an action by an IO, can either file a complaint of criminal defamation or apply to the High Court for an order to file a complaint under sections 212 and 217 of the BNS.

As a result, the system envisioned by the High Court, in addition to having a deterrent effect on the ill-willed complainers, will also have a chilling effect on genuine victims, who might fear going ahead with their case because of the possibility of mechanical prosecution under Sections 212 and 217 of the BNS. While the High Court acknowledges the possibility that the police closure may be ruled incorrect, this is hardly any succour to the ordinary litigant. The logical leap that every closure must mean false prosecution is something that the Court somehow does not seem perturbed by. 

Testing on the Anvil of Practicality

The suggestion, in addition to being tested on the anvil of law, should also be considered in light of the current situation of the Indian criminal system. At first, Umme Farva recommends filing the complaint as soon as a closure report is filed (para 31). However, without qualifying the closure report with “false case”, it also covers other types, such as “mistake of law” and “insufficient evidence”. Thus, these other types unravel the fact that mere filing of a closure report does not automatically imply that the original complaint was legally ‘false’ or malicious. To illustrate, assume police file a closure report, concluding that the dispute is civil in nature. The automatic classification of the original complaint as ‘false information’ under the complaint is legally unsound, especially when it compels the police to criminalise a jurisdictional misunderstanding. Therefore, by omitting classification, the direction facilitates frivolous litigation and frustrates the objective of section 215(1), BNSS, as noted in Mata Bhikh.

Furthermore, by creating a structure of mandatory prosecution, even in the case of a “false case” closure report, the possibility of petty prosecution and hence, violation of the well-accepted principle of de minimis non curat lex cannot be ignored. Moreover, this narrow category fails to create a distinction between genuine false cases and cases categorised as false cases because of a politically influenced investigation. This type of investigation is not imaginative but recognised and deprecated by the Supreme Court in Prakash Singh v. Union of India. Moreover, even if not politically influenced, the investigation may be faulty, as noted in Kailash Gour v. State of Assam. The Court therein ruled that “the benefit arising from any such faulty investigation ought to go to the accused and not to the prosecution” (para 29). While the benefit to the accused cannot be curtailed, the burden cannot be shifted to the victim either, as it will have a chilling effect on genuine victims.

Even if these possibilities are foreclosed, this option of mandate will divert the focus of the already understaffed police force from investigating serious offences to procedural filing of mechanical complaints. Moreover, the state’s prosecutorial team is also diverted towards prosecuting these complainants. It can also not be understated that this mechanisation will add to the already humongous pendency and therefore, add undue pressure to the already overloaded Indian judicial system (see 239th Law Commission Report). Thus, the interests of victims, especially of serious crimes, will be hampered, frustrating the objective of criminal law.

Conclusion

The directives in Umme Farva, which mandate filing complaints for false information alongside closure reports, face significant legal and practical challenges. Legally, Supreme Court precedents like Daulat Ram establish that such action should remain discretionary and requires proof of criminal intent rather than being an automatic procedure. Practically, a mandatory approach fails to distinguish between malicious falsehoods and simple mistakes of law or insufficient evidence. Consequently, enforcing this mandate would likely overburden the judicial system and distract police from investigating serious crimes. The existing system of prosecutorial discretion operates efficiently and effectively within the bounds of law and therefore requires no modification.

The Prior Antecedents Practice Directions on Bail - An Unfair Burden

A specific instance of a person who lied about his previous involvements in crime to get bail comes to the Supreme Court. Going by available empirical metrics, the person is by no means an ordinary accused who finds himself entangled in the Indian criminal justice system. I say this, because most accused persons are from socio-economically marginalised sections of society, whereas the person here claimed to be a lawyer and the head of educational institutions. 

Now, upon seeing that this extraordinary accused person—one with a lot of social and political capital, we can assume—has gamed the system and suppressed facts to get bail, any right-thinking person would be angry, even enraged. But should the sharp practices resorted to by this specific accused person, become the basis either for making observations or passing directions, that would apply to the entire gamut of accused and convicted persons who turn to courts for bail? Probably not. As bad a case of inductive reasoning as any you might see used by partisan policymakers. 

Yet, this is exactly what the Supreme Court recently did in Zeba Khan v. State of Uttar Pradesh [Crl. A. No. 825/2026 decided on 11.02.2026]. The Petitioner had challenged the bail granted to her brother-in-law by the High Court. One of the grounds for challenging the bail order was that this accused person, facing allegations of forgery and fraudulent representations, had lied about his prior criminal history before the High Court. No discussion appears to have been made in the Supreme Court's judgment as to what, if any, steps were taken by the prosecution or Petitioner in highlighting these suppressions before the High Court itself. Nevertheless, upon seeing such conduct, the Supreme Court used the chance to reiterate a feeling that the conduct of this specific accused reflected a "growing and disturbing trend of accused persons securing discretionary relief by suppressing material facts" (Para 43). The answer for the Court was to issue sweeping directions, that an accused person was under "an obligation to disclose all material particulars, including criminal antecedents and the existence of any coercive processes such as issuance of non-bailable warrants, declaration as proclaimed offender, or similar proceedings, duly supported by an affidavit, so as to promote uniformity, transparency and integrity in bail adjudication." (Para 48). An illustrative framework for disclosure was also given (Para 49), along with directions to all High Courts to make necessary rules for their respective areas.

Nobody can have any problem with the idea that courts should not be hoodwinked by a party seeking reliefs. But there are many pathways to achieve this objective of securing "transparency and integrity in bail adjudication", as the Supreme Court puts it. Sadly, the pathway adopted by the Court in Zeba Khan is arguably the worst of the lot. It places an impossibly onerous obligation upon most accused persons, and is likely to delay adjudication of bail petitions and prolong custody, proving that the medicine is definitely worse than the disease.

Firstly, let us be clear that the directions in Zeba Khan are unprecedented. The Supreme Court cites two decisions in its discussion on this aspect. Neither took this drastic a step. The first case, Kushal Duruka, required accused persons to disclose prior bail applications filed in a specific case. The second case, Kaushal Singh, referred to the Punjab & Haryana High Court's Rules which required bail petitions to detail the antecedents of an accused. Notably, the Rules also state that details are not given, an appropriate noting must be made for the bench alerting it to this absence of data. Thus, at no point had any judgment, rule, or guideline, required accused persons to swear an affidavit as to their prior antecedents as a condition for moving a plea for bail. 

Since nobody apart from defence lawyers and judges care about the rights of accused persons, surely many people would have had this idea of demanding an affidavit of clean pasts. Why not do it then? Because it just doesn't work well enough given how the criminal justice system is structured. Discounting the fact of the accused persons who knowingly cheat the system, there are far more persons who simply have absolutely no idea about the specifics of their 'criminal history' as the law calls it. Whatever affidavit they may give may be incomplete. In many cases, since the accused is in custody making filing an affidavit quite difficult and time-consuming, the affidavit may be filed by their next-friend who will know even lesser. What happens in that case: Does a court assume bad faith, or not?  

That accused persons are simply in the dark is down to a mixture of the fact that most accused persons are disadvantaged and disenfranchised, and the fact that the criminal justice system is premised upon a notoriously unfair balance of power where the State holds all the cards when it comes to information. It is not mere rhetoric, but a matter of rule that accused persons are not entitled to copies of even an FIR without running through hoops. They do not get a copy or alert about warrants or other coercive processes, let alone case papers which are kept with lawyers who are frequently changed during the life of a case. 

Compare the ordinary, hapless, accused, with the leviathan which is the State in this scenario. A power to forcefully obtain biographic details about persons accused in offences has existed with the Indian State for more than a century. The most recent statutory version of this power, seen in the Criminal Procedure Identification Act 2022, provides for taking and retaining data for up to 75 years. All this data is then synchronised digitally using a system called the CCTNS. Which leads me to state the obvious: It is a lot easier for the State rather than an accused to provide details about criminal history in 21st Century India.       

For all these reasons, I am highly skeptical that the system will make any great strides towards ensuring either greater transparency or integrity in bail adjudication by this affidavit of past antecedents being filed by accused persons. What is much more likely, as is already being seen in Delhi, is that bail applications will take longer to get heard. The courts may ask police to now verify details in this affidavit — in effect, unnecessarily wasting everyone's time because it is something that police can do at the outset. Then, if there are discrepancies, which there will be in many cases, it will give the prosecution more opportunity for time-wasting and distractions by alleging bad faith on part of an accused, requiring a court to spend time on this issue rather than the primary issue of whether a person should be in custody on allegations of having committed a crime.

The real question to probe in Zeba Khan was why did the existence of such a notorious rapsheet go unnoticed by the prosecution before all courts, and then formulating guidelines which placed a duty upon the state agencies to undertake a verification of criminal antecedents in bail cases. A step like would meet the desired objective, in a fair and reasonable manner. Instead, as has been often the case, the legal system has simply erected fresh obstacles for persons trying to restore their liberty, while making it easier for the State to continue their incarceration by default.