The research, conducted by Kavali Munisekhar and Bangalore Morarji at Vellore Institute of Technology, develops the Human-Centred Constitutional Urban AI Governance Framework and operationalises it as the Constitutional Smart-City Governance Index, or CSGI. The instrument scores urban AI against Article 21 of the Indian Constitution, human-centred AI principles, and Sustainable Development Goals 11 and 16. Six pillars—constitutional rights, human-centred AI, sustainable development, service delivery, participation, and accountability—comprise thirty-eight indicators with operational definitions, official secondary-data sources, a transparent normalisation-and-weighting model, and a validation strategy. The framework is deliberately positioned at the intersection of urban studies, public administration, law, and AI governance: the Constitution supplies the normative foundation, but the analytical machinery is that of governance research rather than doctrinal legal commentary.

What makes Article 21 an unusually rich constitutional resource for exactly this task is five decades of Supreme Court jurisprudence that have expanded it from a procedural shield into a substantive charter of dignified urban life. Shelter and livelihood were recognised in Olga Tellis v. Bombay Municipal Corporation in 1985, dignity and the bare necessities of life in Francis Coralie Mullin in 1981, a clean environment in Subhash Kumar v. State of Bihar in 1991, and, decisively for the digital era, informational privacy subject to a structured proportionality test in Justice K.S. Puttaswamy v. Union of India in 2017. When an integrated command and control centre fuses surveillance feeds, when a welfare platform authenticates biometric identity before releasing entitlements, or when a predictive model allocates sanitation resources across neighbourhoods, Article 21 is not a peripheral legal technicality; it is the constitutive standard of legitimate state action.

The CSGI was applied to four leading Indian smart cities—Bengaluru, Pune, Surat, and Indore—selected for their AI maturity, domain diversity, and prominence within national urban policy. Bengaluru’s portfolio is dominated by mobility and public safety, including adaptive sensor-driven traffic signal control and an extensive Safe City surveillance programme integrating video analytics with policing command centres. Pune centres on an integrated command centre coordinating traffic, emergency, and utility operations alongside platform-based citizen engagement channels. Surat emphasises hydrological monitoring and flood early warning linked to upstream dam management, while Indore exemplifies the operational-excellence pathway built on comprehensively re-engineered waste management with GPS-tracked collection fleets and monitored processing facilities.

The governance-opacity audit that followed produced the study’s most striking finding. The four cities are 86% to 93% opaque on the transparency and accountability artifacts a rights-based assessment requires. No city publishes a data-protection impact assessment for its AI or surveillance systems, none maintains a public register of deployed algorithms, none has adopted a municipal AI-governance policy, and none subjects its algorithmic systems to independent audit—each of these four artifacts is absent in 100% of the cases examined. Only integrated command centre operational reporting and special purpose vehicle-to-council reporting exist in even rudimentary form, and only partially. Bengaluru, the city that tops the national liveability ranking, is on this measure the most opaque of the four. The finding reframes the study’s data scarcity: the missing values on the rights and accountability pillars are not a gap in the researcher’s dataset but a measurement of a structural condition—Indian urban AI is being deployed faster than the disclosure and audit infrastructure that would make it accountable under Article 21.

The index-divergence analysis then exposed what the researchers call a constitutional blind spot in India’s urban metrics. The CSGI correlates strongly negatively with the Ease of Living Index (Spearman ρ = −0.80) yet positively with the Municipal Performance Index (ρ = +0.80) and the Swachh Survekshan cleanliness survey (ρ = +0.95). Bengaluru crystallises the divergence: first on liveability, it ranks last on municipal performance, last on cleanliness, and last on the CSGI. The Ease of Living Index, in turn, is negatively concordant with every other instrument. When such a metric conditions reputational and increasingly fiscal rewards, it steers cities toward the dimensions it can see and away from the rights and accountability dimensions it cannot. An incentive that is blind to rights and accountability will predictably produce cities that are optimised for visible liveability and hollow on constitutional substance.

A Monte Carlo weight-robustness test addressed the objection that composite index rankings are artefacts of the analyst’s weights. The CSGI was recomputed for 20,000 pillar-weight vectors drawn uniformly from the weight simplex, so that every admissible weighting—including schemes that all but ignore the constitutional pillars—is equally represented. Bengaluru ranks last under 100% of the sampled weightings: there exists no admissible combination of pillar weights on which it escapes the bottom position. Its entire CSGI distribution (90% interval 1.55–2.50) lies below those of the other three cities. The ordering among the top three is genuinely weight-dependent—Surat leads under 45% of weightings, Pune under 40%, and Indore under 15%, with none ever ranking last. The substantive claim the paper rests on is therefore the robust one: whatever normative weight a reader assigns to constitutional rights relative to efficiency, the city that leads the national liveability ranking occupies the lowest position on a rights-inclusive assessment across the entire weight space.

The pillar-by-pillar evidence reveals an inverted governance profile: every city scores highest on the outcome pillars—sustainable development and service delivery—and lowest on the normative and institutional pillars of constitutional rights protection and institutional accountability. Indian urban AI, on this evidence, is operationally capable and constitutionally under-governed. Surat’s flood early-warning and utility analytics constitute a substantive resilience achievement aligned with SDG 11.5 and protective of life in the literal Article 21 sense, while Indore’s waste system delivers measurable SDG 11.6 outcomes. Yet the distributional caveat is systematic: area-based development concentrated smart infrastructure in selected zones, and data coverage of informal settlements lags, so gains are unevenly experienced. Human oversight is strongest where operational stakes enforce it—Surat’s flood early-warning embeds forecasts in an emergency-management chain in which officials interpret model outputs, exemplifying the high-automation/high-control combination that human-centred AI prescribes—but transparency, explainability, and independent oversight are largely absent.

The paper’s central interpretive claim is that constitutional safeguards should be understood as infrastructure for sustainable urban AI, not as external constraint upon it. Doctrinally, Article 21 already contains the evaluative materials—dignity, livelihood, environment, privacy, fair procedure—that AI ethics frameworks attempt to supply from softer sources; anchoring urban AI governance in Article 21 converts persuasive principles into enforceable standards with an existing remedial apparatus. Institutionally, constitutional anchoring supplies what the global AI-principles consensus lacks: a domestic adjudicator, a body of interpretive method in the proportionality test, and legitimacy independent of technology vendors and executive discretion. Instrumentally, the sustainability evidence suggests constitutional compliance and SDG performance are complements, not substitutes: Surat’s life-protective early warning and Indore’s environmental gains are advances in Article 21’s substantive guarantees, and the trust generated by rights-respecting design is itself a resource for the sustained citizen cooperation on which sanitation, resilience, and mobility systems depend.

The study offers tiered recommendations sequenced from disclosure through assessment to independent oversight, designed to be adoptable within existing legal authority. The Government of India should condition central assistance for AI-enabled urban projects on a published Constitutional and Data-Protection Impact Assessment applying the Puttaswamy proportionality test, and task the IndiaAI Mission’s governance apparatus with a national public register of urban algorithmic systems. The Ministry of Housing and Urban Affairs should extend the DataSmart Cities Strategy into a Model Municipal AI Governance Policy with procurement clauses for explainability, bias testing, audit rights, and exit provisions, and rebalance mission-evaluation instruments so rankings reward published impact assessments, deliberative participation, and functioning algorithmic grievance routes alongside output metrics. State governments should notify municipal AI oversight provisions under state municipal acts, including council approval for surveillance-capable deployments. Urban local bodies should institutionalise upstream participation through ward-committee deliberation on proposed deployments before procurement, and extend grievance platforms with an algorithmic-decision category carrying defined timelines and a human-review guarantee. Smart City special purpose vehicles should report AI deployments, incidents, and audit findings to the council on a fixed calendar and publish system inventories in open formats. The Data Protection Board should prioritise urban public-sector deployments and publish reasoned decisions, while the Comptroller and Auditor General should pilot performance-audit methodology for urban algorithmic systems. The researchers acknowledge that the present application is an illustrative pilot rather than a validated measurement—because comparable cross-city data remain sparse, the data-scarcity itself is a substantive finding and a precondition the field must remedy before a fully quantitative CSGI can be validated. But the diagnosis is unambiguous: the right to life is being advanced in outcome and exposed in process, and the exposure is systemic rather than city-specific. Cities that govern AI through their constitutional commitments will not merely avoid harm; they will convert the Constitution’s oldest promise, the protection of life, into the organising purpose of their newest infrastructure.

Subject of Research: AI-enabled smart cities and the constitutional right to life in urban India

Article Title: Can AI-enabled smart cities advance the constitutional right to life? A rights-based index, governance-opacity audit, and robustness analysis for sustainable urban India

Article References: Munisekhar, K., & Morarji, B. (2026). Can AI-enabled smart cities advance the constitutional right to life? A rights-based index, governance-opacity audit, and robustness analysis for sustainable urban India. Frontiers in Sustainable Cities, 8, Article 1959012. https://doi.org/10.3389/frsc.2026.1959012

Image Credits: AI Generated

DOI: 10.3389/frsc.2026.1959012

Keywords: artificial intelligence, smart cities, constitutional governance, human-centred AI, urban India, SDG 11, digital governance, AI-enabled, smart, cities, advance, constitutional

The appeal of AI in this setting rests on a simple asymmetry. Clinical trials generate enormous volumes of structured and unstructured data—imaging, laboratory values, genomic profiles, clinician notes, wearable-device streams and patient-reported outcomes—yet historically only a small fraction of that information has been used systematically in trial design and conduct. Machine-learning models, particularly modern deep-learning and foundation-model architectures, excel at extracting signal from exactly these high-dimensional, heterogeneous datasets. When trained on curated clinical data, they can identify which patient subpopulations are most likely to respond to a therapeutic candidate, predict which sites will struggle with recruitment, flag data inconsistencies long before database lock, and estimate the probability that an ongoing trial will meet its primary endpoint. The result is a shift from retrospective, intuition-driven trial management toward a prospectively optimised, continuously adaptive process.

Patient recruitment illustrates the potential most vividly. Failed recruitment and poor participant retention are among the leading reasons that trials miss their timelines or terminate early. Traditional approaches rely on broad eligibility criteria, manual chart review and outreach from a limited number of academic centres, which systematically under-enrols patients from rural areas, lower-income communities and historically marginalised groups. AI-driven approaches invert this model. Natural-language processing systems can scan millions of de-identified electronic health records to locate every patient matching a complex eligibility profile, including patients whose relevant conditions are described only in free-text notes rather than coded diagnoses. Matched patients can then be connected with trial sites through their treating physicians or through decentralised trial infrastructure that brings the study to the participant rather than the reverse. When designed carefully, such tools can also be tuned to improve diversity in enrolment, addressing a long-standing scientific weakness as well as an ethical one: a trial population that does not resemble the eventual treated population limits the generalisability of the results.

Trial design itself is being recomputed. Classical phase II and phase III trials typically use fixed designs conceived months before enrolment begins, and any mid-course revision requires protocol amendments that can delay readouts by many months. AI-assisted design tools draw on large repositories of historical trial outcomes, natural-history data and simulation frameworks to model, before a single patient is enrolled, how a trial will behave under different assumptions about effect size, dropout, endpoint variability and site performance. Bayesian adaptive designs, which allow the randomisation ratio and dose allocation to shift as interim data accumulate, become far more practical when machine-learning models supply reliable predictive components. Digital twins—computational replicas of individual patients built from rich baseline data—offer an emerging complement, allowing a portion of the control information in a trial to be estimated rather than observed, and thereby reducing the number of real participants who must be randomised to achieve a given statistical power. Regulators have begun engaging with these methods, and several external control arms constructed from historical or registry data have already supported regulatory submissions in rare-disease oncology.

Once a trial is running, AI changes the economics of monitoring. Risk-based monitoring, in which scrutiny is concentrated on the sites and data elements most likely to harbour errors, depends on recognising patterns across thousands of simultaneous data streams—precisely the task that anomaly-detection algorithms perform well. Centralised statistical monitoring can flag sites whose data distributions deviate from the norm, whether because of fraud, systematic measurement error or simple process breakdown, without the cost of sending monitors to every site on a fixed schedule. Automated adverse-event signal detection can surface safety trends earlier in the data stream, shortening the interval between an emerging risk and a protocol response. Speech-recognition and summarisation models are increasingly used to draft clinical notes, assist with adjudication of endpoints that require expert review, and reduce the administrative burden that currently consumes a large share of investigator time. Because the marginal cost of applying a trained model to new data is close to zero, these efficiencies can scale across an entire portfolio rather than applying to a single study.

The pharmaceutical industry’s interest follows directly from the arithmetic. Industry analyses have repeatedly estimated that bringing a new drug to market costs on the order of one to two billion dollars, with clinical development accounting for the majority of that expenditure and with most of the cost attributable to failed trials. Even a modest improvement in the probability of technical success, achieved through better target-patient matching or earlier detection of futility, translates into hundreds of millions of dollars in expected savings per programme and, more importantly, into faster access for patients to therapies that work. Every month shaved from a development timeline extends the effective patent-protected market life of a medicine, which strengthens the commercial case, but the public-health case is at least as strong: pipelines that iterate faster can respond more quickly to emerging pathogens, to rare diseases that currently have no treatment at all, and to the individualisation of therapy in fields such as oncology where one-size-fits-all efficacy is the exception rather than the rule.

None of this means the transformation is automatic. Machine-learning models inherit the biases of the data on which they are trained, and clinical datasets under-represent precisely the populations in whom trial evidence is weakest. An algorithm trained mostly on data from large academic hospitals may perform poorly for patients managed in community settings, and a recruitment tool optimised narrowly for speed could worsen enrolment diversity if fairness constraints are not built in explicitly. The opacity of complex models also sits awkwardly with regulatory expectations: a sponsor that uses an AI-derived covariate in a statistical analysis plan, or an AI-curated external control arm, must be able to explain to reviewers how the model was built, validated and monitored. Regulators including the United States Food and Drug Administration and the European Medicines Agency have signalled that they will evaluate such tools under existing frameworks for risk-based software validation, but the guidance landscape is still maturing, and sponsors who treat AI components as unexamined black boxes do so at their own regulatory peril.

Data governance presents a second structural challenge. The models that promise the greatest gains in recruitment and design are those trained on the largest and most diverse clinical datasets, yet health data are fragmented across institutions, jurisdictions and incompatible record systems, and privacy law constrains how they can be pooled. Federated learning, in which models are trained across multiple sites without moving the underlying patient data, offers a technically elegant partial solution, but it introduces its own questions about model ownership, auditability and the equitable distribution of the value created. Standard-setting efforts, from common data models to documented provenance for training sets, will determine whether AI-enabled trials become a broadly shared capability or a competitive advantage concentrated in a handful of organisations with the largest proprietary data estates.

The most realistic near-term picture is therefore not one of autonomous AI running trials, but of a human-machine division of labour in which algorithms perform the enumeration, matching, monitoring and simulation that humans cannot do at scale, while clinicians, statisticians and regulators retain judgment over what counts as evidence. Under that division of labour, the measurable signs of change are already visible: screening times measured in days rather than months at leading sponsors, growing numbers of adaptive and model-informed designs entering regulatory review, and decentralised, data-rich trial formats that were logistically implausible a decade ago. If those trends continue, the defining feature of the next generation of clinical trials will not be any single algorithm but the integration of computation into every stage of the evidentiary pipeline—from the first patient matching query to the final submission—producing trials that are faster, smaller where possible, larger where necessary, and ultimately more representative of the patients the resulting medicines are meant to serve.

Subject of Research: AI-enabled clinical trials

Article Title: AI-enabled clinical trials

Article References: Raynaud, M., Trayanova, N., Mannon, R. B., André, F., Doraiswamy, P. M., & Loupy, A. (2026). AI-enabled clinical trials. Nature Reviews Bioengineering. https://doi.org/10.1038/s44222-026-00487-7

Image Credits: AI Generated

DOI: 10.1038/s44222-026-00487-7

Keywords: AI-enabled, clinical, trials, scientific research, peer-reviewed research, research findings