Differential Bail in the Delhi Riots UAPA Case: Supreme Court’s Evolving Approach

**Introduction**

The Supreme Court of India’s January 2026 decision in *Gulfisha Fatima v State (NCT of Delhi)*, 2026 SCC OnLine SC 10, marks a significant development in the law governing bail under the Unlawful Activities (Prevention) Act 1967 (UAPA). Arising out of the 2020 Delhi riots “larger conspiracy” prosecutions, the Court declined bail to Umar Khalid and Sharjeel Imam while granting conditional bail to five co-accused: Shifa Ur Rehman, Mohd. Saleem Khan, Meeran Haider, Shadab Ahmed, and Gulfisha Fatima. The judgment has renewed debate about how courts should apply the stringent “prima facie true” standard under section 43D(5) UAPA, and how far individual roles can, and should, be differentiated in conspiracy cases. This post examines the legal background, the Court’s reasoning, and its broader implications for terrorism-related prosecutions and personal liberty.

**Legal Background**

UAPA is India’s principal anti-terror legislation, designed to address activities threatening the sovereignty and integrity of the State. Section 43D(5) creates an exceptional bail regime: where a person is accused of offences under Chapters IV and VI (terrorist acts and terrorist organisations), the court must refuse bail if, on a perusal of the case diary or final report, there are “reasonable grounds for believing that the accusation against such person is prima facie true.” This is a marked departure from the ordinary CrPC standard, where gravity of offence and risk factors are balanced against the presumption of innocence and the right to personal liberty.

The Supreme Court’s leading authority on section 43D(5) is *National Investigation Agency v Zahoor Ahmad Shah Watali* (2019) 5 SCC 1. In *Watali*, the Court held that, at the bail stage, the court is not to conduct a mini-trial or weigh evidence for admissibility or probative value. Instead, it must accept the prosecution’s materials at face value and ask whether they disclose a prima facie true case. This highly prosecution-friendly standard has been followed in numerous UAPA decisions, including Delhi High Court rulings such as *Jamsheed Zahoor Paul v State (NCT of Delhi)* (2024) and *Ammar Abdul Rahiman v NIA* (2024), where courts emphasised that once the section 43D(5) threshold is crossed, bail is to be denied “as a rule”.

At the same time, constitutional principles of personal liberty under Article 21, and comparative jurisprudence on terrorism-related detention (for example, UK decisions on control orders post-*A v Secretary of State for the Home Department* [2004] UKHL 56 and ECHR case law under Article 5), have underscored that special regimes cannot wholly extinguish judicial scrutiny. Indian courts have therefore been slowly recalibrating the rigidity of *Watali* by underlining the need to examine the nature of an accused’s specific role and the proximity of alleged conduct to acts of violence.

**Critical Analysis**

In *Gulfisha Fatima*, the Supreme Court carefully parsed the role-attribution across different accused persons. According to publicly available reports, the Court found that there was prima facie material suggesting a “central and formative role” for Umar Khalid and Sharjeel Imam in the alleged conspiracy. By contrast, for the five co-accused who were granted bail, the Court held that the prosecution’s materials, when read as a whole, did not justify continued incarceration in the face of prolonged trial and more peripheral roles. It is important to stress that the Court did not depart from the *Watali* framework; rather, it applied it with a sharper focus on differentiation within a conspiracy.

Under Indian conspiracy law (Penal Code 1860, section 120B), and by analogy with UK and Commonwealth authorities, mere association or ideological sympathy is insufficient: there must be agreement to commit an unlawful act or a lawful act by unlawful means. In complex prosecutions involving protests, political speech and civil society mobilisation, the line between legitimate dissent and criminal conspiracy is often contested. The Delhi riots prosecutions have attracted criticism, including concerns about over-breadth of the alleged “larger conspiracy” and reliance on protected speech.

Within this fraught context, the Supreme Court’s approach can be seen as an attempt to reassert individualised assessment while staying within the strict statutory bail bar. For Khalid and Imam, the Court appears to have accepted that the prosecution materials—speeches, communications, and alleged coordination—went beyond political rhetoric and, taken at face value, pointed to active planning and instigation amounting to a central conspiratorial role. Applying *Watali*, the Court thus held that the accusations against them were prima facie true and that section 43D(5) mandated refusal of bail.

For the co-accused, the Court appears to have considered several factors that, read together, tilted the balance in favour of bail: the nature of their alleged participation (for example, local mobilisation, meetings, or membership of organisations rather than command-level decision-making); the absence of direct material linking them to instigation or planning of specific episodes of violence; the stage and likely duration of the trial; and their period of pre-trial custody already undergone. This resonates with a more nuanced reading of *Watali*, where “prima facie true” must still be established in relation to each accused, not the conspiracy in the abstract.

Human rights concerns are central to this analysis. Prolonged pre-trial detention under a stringent bail regime can, in effect, amount to punishment before conviction, contrary to the presumption of innocence. The Supreme Court has, in other contexts, emphasised the dangers of “trial by incarceration”, notably in *Union of India v K.A. Najeeb* (2021) 3 SCC 713, where it granted bail despite UAPA charges due to excessive delay. Although *Najeeb* was not reported as being central in *Gulfisha Fatima*, the spirit of that decision—recognising constitutional courts’ power to secure liberty where statutory bars produce disproportionate detention—arguably informs the Court’s willingness to grant bail to less centrally-placed accused.

Comparative jurisprudence from the UK and European Court of Human Rights also illustrates a similar trend: special terrorism regimes are tolerated, but courts increasingly require specificity in the evidence connecting each suspect to the risk justifying continued detention. The *Gulfisha Fatima* ruling reflects this movement, even as it continues to give substantial deference to prosecutorial narratives at the bail stage for those alleged to be architects of the conspiracy.

**Opinion & Outlook**

From a doctrinal standpoint, *Gulfisha Fatima* does not overturn *Watali*, but it subtly recalibrates its operational impact. The message to trial courts and investigating agencies is twofold. First, section 43D(5) remains a powerful statutory shield against bail in serious UAPA cases, particularly where there is material pointing to central planning, coordination, or ideological leadership of violent acts. Secondly, however, an undifferentiated approach that treats every participant in a large-scale protest movement or loosely defined network as equally culpable will not automatically survive scrutiny at the apex level.

Going forward, defence counsel are likely to rely on this decision to argue for a more granular evaluation of their clients’ roles, especially where the evidence consists largely of presence at protests, organisational membership, or political speech devoid of a clear nexus to violence. Prosecutors, on the other hand, may respond by seeking to develop more specific evidentiary chains—digital communications, financial trails, and witness statements—that position particular accused as critical nodes in the conspiracy.

The decision also interacts with broader debates on reforming UAPA’s bail provisions. Critics have argued that the “prima facie true” standard, as interpreted in *Watali*, pushes the system towards pre-conviction punishment. While legislative change appears unlikely in the short term, judgments like *Gulfisha Fatima* and *K.A. Najeeb* demonstrate that constitutional courts can mitigate the harshest effects by emphasising role-based scrutiny and the temporal dimension of pre-trial custody.

In terms of precedent, *Gulfisha Fatima* will likely be cited in future bail applications to justify both outcomes: denial of bail where central leadership and active planning are alleged, and grant of bail where the accused’s role is peripheral, the link to violence is tenuous, or delay becomes oppressive. Its true significance may therefore lie less in doctrinal innovation and more in rebalancing a jurisprudence that had, post-*Watali*, become heavily weighted in favour of automatic denial.

**Conclusion**

The Supreme Court’s ruling in *Gulfisha Fatima v State (NCT of Delhi)* represents an incremental, yet important, refinement of UAPA bail jurisprudence. By distinguishing between central and peripheral roles within an alleged conspiracy, the Court reaffirms that even in terrorism-related cases, individualised assessment, proportionality, and the overarching guarantee of personal liberty remain vital. While the high threshold of section 43D(5) endures for those alleged to be principal architects of violence, the decision opens space for more nuanced bail determinations in complex, politically charged prosecutions—an evolution that will shape the landscape of anti-terror trials in the coming years.