How NHAI’s AI‑Driven Contract Review May Prompt Judicial Scrutiny of Administrative Power and Natural‑Justice Duties

The launch of Margsarthi by the National Highways Authority of India represents a groundbreaking application of artificial intelligence to the analysis of contract documents, wherein the system automatically identifies defects, contradictions, and potential dispute risks, thereby transforming the way contract documents are reviewed; according to the announcement, the tool significantly reduces the time required for document review, shrinking a process that previously required multiple days to an operation completed within minutes, thereby delivering unprecedented efficiency in contractual oversight; the capability of Margsarthi to flag defects and contradictions is portrayed as a means to enhance the quality of contract administration, potentially allowing the authority to address issues before they evolve into formal disputes; by employing this technology, the authority aims to streamline its contract management processes, which historically involved extensive manual scrutiny, and to align its operational practices with modern digital innovations; the introduction of such an AI‑driven review mechanism raises questions concerning the procedural implications for contractors whose agreements are subject to automated assessment, including considerations of transparency, fairness, and the avenue for remedies should the AI flag errors; stakeholders anticipate that the reduction in review time may accelerate project execution on national highways, yet the legal community may scrutinize how the authority ensures that the AI’s determinations are consistent with its statutory duties and established principles of natural justice.

One question is whether the National Highways Authority of India possesses the statutory power to incorporate an artificial‑intelligence system such as Margsarthi into its contract‑administration functions, given that its legislative mandate may traditionally envisage manual review processes and the delegation of decision‑making to human officials, and if the statutory scheme permits the authority to modernize its processes, it must nonetheless ensure that any delegation to an artificial‑intelligence system does not abdicate the duty to make determinations that are legally defensible and subject to oversight.

Perhaps the more important legal issue is whether the adoption of an automated tool alters the nature of the authority’s administrative action to the extent that the resulting determinations become subject to the same standards of reasoned decision‑making and procedural safeguards that apply to conventional administrative orders.

The answer may depend on whether contractors affected by AI‑generated defect flags are afforded a meaningful opportunity to be heard before any adverse contractual consequence is imposed, thereby engaging the principles of natural justice that require notice and a chance to contest findings, and moreover, the right to contest an AI‑generated defect flag may invoke the doctrine of audi alteram partem, requiring that the authority disclose the specific defects identified and the criteria applied, thereby enabling effective participation.

Perhaps the procedural significance lies in whether the authority must provide a detailed explanation of the AI’s reasoning, because without such transparency the affected parties may be unable to formulate an effective response, potentially rendering the process arbitrarily opaque, and in addition, the potential for algorithmic bias necessitates that the authority conduct regular audits of the AI model to ascertain that it does not disproportionately disadvantage particular categories of contractors, aligning with the constitutional guarantee of equality before the law.

Another possible view is that the use of an algorithmic system raises questions of transparency and accountability, since the legal regime governing administrative action typically obliges the decision‑maker to disclose the material basis for its conclusions, and the opacity of machine‑learning models may frustrate that requirement, and perhaps a court would examine whether the authority has instituted adequate safeguards, such as human oversight or validation of AI outputs, to mitigate the risk of erroneous or biased determinations that could impinge upon contractual rights.

The issue may require clarification from the authority regarding the remedial avenues available to contractors who challenge an AI‑flagged defect, including whether an internal appeal, an external review, or a writ petition in a higher court constitutes the appropriate mechanism for redress, and should the authority fail to provide a clear procedural roadmap for challenges, affected contractors might resort to filing a writ of certiorari, arguing that the AI‑driven process violates the principles of natural justice and statutory mandates.

Perhaps the legal position would turn on whether the AI’s findings are treated as substantive determinations that bind the parties or merely as advisory inputs subject to further human discretion, a distinction that directly influences the scope of judicial review, and a fuller legal conclusion would require an assessment of whether existing regulatory frameworks governing public‑sector procurement and digital governance contain provisions that specifically address the deployment of artificial‑intelligence tools, and whether the authority must develop additional policies to ensure compliance with overarching principles of fairness, non‑discrimination, and statutory duty.

Thus, the introduction of Margsarthi may serve as a catalyst for broader judicial scrutiny of governmental adoption of advanced technologies, prompting courts to articulate standards that balance efficiency gains with the preservation of procedural rights and accountability in administrative decision‑making, and consequently, legislative bodies may consider enacting specific provisions that delineate the permissible scope of artificial‑intelligence deployment in public administration, thereby furnishing courts with a statutory touchstone for evaluating compliance and safeguarding rights.