To Regulate or Not to Regulate: India's AI Dilemma

The proposed consultation paper is a government document seeking views on how India should regulate artificial intelligence. Its importance lies in clarifying an approach that has repeatedly shifted between light-touch governance and stronger intervention. It may address which risks need rules and which regulatory tools are suitable.

The announcement followed several years of policy development. India has considered responsible-AI principles, risk-based oversight, high-risk systems, algorithmic accountability, and safeguards against bias, discrimination, deepfakes, and unlawful content. A March 2024 advisory showed how quickly the government could intervene, and its revision exposed concerns about legal certainty.

The article does not provide the paper’s final contents. It does establish the central questions: whether existing laws are enough, what additional safeguards are needed, and when dedicated legislation becomes necessary. The consultation could therefore make India’s regulatory direction more predictable while preserving room for innovation.

AI regulation is the use of laws, official rules, standards, and oversight to control how AI is developed and used. It aims to reduce harms such as discrimination, privacy violations, unsafe outputs, and weak accountability. The article presents regulation as a way to balance public safety and innovation.

Voluntary industry guidelines work differently. They ask companies to follow principles such as fairness, transparency, and accountability without necessarily creating legally enforceable duties. India’s MeitY committee recommended flexible self-regulatory mechanisms, while NITI Aayog promoted responsible-AI principles and risk-based oversight. These measures can guide behaviour but do not equal a comprehensive statutory framework.

India’s approach has included both voluntary and mandatory elements. Existing laws and executive advisories can constrain conduct, but the March 2024 advisory raised questions about legal authority and predictability. The proposed consultation may clarify whether stronger, enforceable AI-specific rules are needed.

The article describes a broad, application-focused approach rather than one limited to advanced foundation models. India has discussed high-risk AI systems, algorithmic accountability, and user harm. This reflects concern about what AI does in society, not only how powerful a model is.

The clearest example is MeitY’s revised 15 March 2024 advisory. It retained provisions on unlawful content, bias and discrimination, electoral integrity, unreliable AI outputs, and identifying AI-generated misinformation and deepfakes. Earlier policy work also identified data privacy, security, intellectual property, fairness, transparency, and accountability as concerns. These measures target risks through duties, warnings, labelling, and oversight.

The article therefore supports a wide reading of India’s proposed direction. It does not provide a complete list of sector-specific harms or final rules. Instead, it shows a preference for regulating AI applications and associated harms, while leaving the precise legal structure to future consultation and policymaking.

India began with policy frameworks rather than a comprehensive AI statute. The 2018 Task Force and NITI Aayog promoted enabling policies, ethical safeguards, privacy protection, security, and responsible deployment. In April 2023, the government said it was not considering legislation, while still recognising risks including bias, discrimination, privacy, and accountability.

That position was never entirely hands-off. The proposed Digital India Act contemplated high-risk AI regulation and algorithmic accountability. In March 2024, MeitY briefly required permission before some unreliable AI models could be offered in India, then withdrew that requirement after industry criticism. Later, the minister considered legislation for publisher and creator rights and questioned self-regulation.

By October 2024, the emphasis had shifted toward flexible, harm-focused regulation that would preserve innovation. The October 2026 consultation announcement continues this evolution. It could help settle whether India needs targeted laws, stronger oversight under existing laws, or a broader AI-specific framework.

Regulating applications focuses on real-world effects rather than treating every AI system as equally risky. The article links this preference to India’s effort to balance innovation with safety. A broad technology-wide regime could be heavily prescriptive, while application-based rules can concentrate attention on systems that threaten users or society.

India’s proposals illustrate this logic. The Digital India Act contemplated high-risk AI systems, and the government said AI could be regulated through the lens of user harm. The March 2024 advisory targeted unlawful content, bias, discrimination, electoral integrity, unreliable outputs, misinformation, and deepfakes. These are harms connected to particular uses and outcomes.

The approach also resembles the minister’s preferred model for data protection and telecommunications laws, which he described as leaving space for innovation while addressing societal harms. However, application-focused regulation may leave difficult questions about definitions, enforcement, and accountability. The consultation could clarify how India will draw those boundaries.

Existing laws can address some AI harms without waiting for a new statute. Voluntary commitments can also encourage responsible development and preserve flexibility. But the article questions whether these tools are adequate, especially for bias, privacy, accountability, publisher rights, and content risks. Weak or unclear duties may leave users without consistent protection.

The March 2024 advisory shows the trade-off. Its first version required permission for some unreliable AI models, but criticism led MeitY to remove that requirement and compliance reporting. The revised version retained safeguards on unlawful content, discrimination, elections, labelling, misinformation, and deepfakes. The episode also raised concerns about the legal basis and predictability of executive action.

New AI-specific legislation could clarify responsibilities, rights, enforcement, and high-risk uses. It might also increase costs or slow experimentation if overly prescriptive. The article does not predict outcomes, but the proposed consultation could make India’s balance between innovation, safety, and certainty clearer.

AI is not one activity with one type of harm. It can process personal data, influence decisions, generate content, affect elections, create cybersecurity risks, and raise questions about ownership and compensation. These issues already touch different legal and regulatory areas. That makes a single, comprehensive solution difficult to design.

India’s policy history reflects this complexity. NITI Aayog identified ethical concerns, data privacy, security, and intellectual property. MeitY’s work added fairness, transparency, and accountability. The 2024 advisory addressed unlawful content, discrimination, electoral integrity, unreliable outputs, misinformation, and deepfakes. The proposed Digital India Act also contemplated high-risk systems and algorithmic accountability.

This breadth helps explain India’s preference for a techno-legal approach using existing laws and flexible measures. It also explains the continuing debate over AI-specific legislation. A future framework may need shared principles plus rules for particular risks and applications. The article leaves that design question open for consultation.