India plans an AI regulation paper within one month, putting safety, deepfakes and human welfare at the center of a national consultation. Electronics and IT Minister Ashwini Vaishnaw announced the deadline on October 8 while launching the World Bank Group's 2026 report on artificial intelligence.

 

The document has not been released, and no draft legal text is available. Vaishnaw said it would provide the basis for future regulation, while stressing that law alone cannot manage AI risks and that industry must accept primary responsibility.

 

The announced consultation will focus on four areas:

  • Safety in the development and use of AI
  • Deepfakes and other forms of user harm
  • Skills, inclusion and a human-first approach
  • Industry responsibility alongside government rules

 

India Plans AI Regulation Paper Within One Month

Vaishnaw said the government would publish the regulation-focused consultation paper within a month and use the responses to shape its next steps. That timetable points to a public policy process rather than an immediately enforceable AI law.

 

The announcement came at the New Delhi launch of the World Development Report 2026, hosted by the IndiaAI Mission, the Ministry of Electronics and Information Technology and the World Bank Group. The event focused on how developing economies can adopt AI while managing the institutional and social risks that accompany it.

 

India's current approach spreads oversight across existing laws, sector regulators and voluntary or technical safeguards instead of relying on one comprehensive AI statute. The consultation could begin to define where that framework is insufficient, but the minister did not announce a bill, enforcement date or new regulator.

 

Safety, Deepfakes and Human Welfare Lead the Scope

Vaishnaw identified AI safety as the consultation's most important concern. He also highlighted cybersecurity, user harm, psychological disorders and deepfakes, describing manipulated media as a growing social problem because increasingly convincing content can mislead people and distort public narratives.

 

The human-first element gives the planned paper a wider remit than content moderation. Skills development and inclusion are expected to be part of the consultation, reflecting the government's argument that productivity gains should not leave workers or citizens without the capacity to participate.

 

Those priorities align with work already funded through the IndiaAI Mission. Its Safe and Trusted AI pillar supports bias mitigation, algorithm auditing, explainable AI, privacy-preserving techniques and deepfake detection, while the broader mission also invests in compute, datasets, foundation models, skills and startups.

 

The consultation may connect those technical programs to clearer obligations for companies. Vaishnaw said industry must carry the bulk of the responsibility, suggesting that developers and deployers could be expected to combine governance processes with technical controls rather than wait for enforcement after harm occurs.

 

Background Reading

 

Agent Autonomy and Liability Remain Open Questions

Earlier policy discussions have explored consent for synthetic content, limits on autonomous agents, regulatory sandboxes and rules for retaining or reusing data. These subjects are plausible candidates for consultation, but they should not be treated as settled provisions until the government publishes the paper.

 

Agentic systems expose a particularly difficult gap. Traditional intermediary rules were designed for platforms that host or transmit third-party material, while generative systems create, select or transform outputs. That makes responsibility harder to divide among a model provider, application developer, deploying company and end user.

 

The same problem appears when an AI agent acts rather than merely answers. A system may search, communicate, purchase or initiate transactions under a user mandate. Regulation must distinguish a model error from an unauthorized action, decide which logs are needed for accountability and clarify who pays when automated decisions cause loss.

 

Deepfake controls raise a different set of tradeoffs. Consent signals and provenance tools can help identify synthetic media, but effective rules must also address impersonation, fraud and distribution without suppressing legitimate satire, accessibility tools or creative work. The consultation process gives affected groups a chance to test those boundaries before legislation is drafted.

 

Consultation Marks a Shift From Existing-Law Approach

Indian officials have previously emphasized adapting existing law and encouraging voluntary compliance as AI capabilities evolve. A dedicated regulation paper signals that the government now sees value in a broader framework, even if the final system remains sector-based and combines legal duties with technical standards.

 

The shift does not guarantee a single AI act. India could instead amend current digital rules, issue sector-specific requirements or create shared obligations for high-risk uses. The published paper will need to identify which institutions would enforce the framework and how its requirements interact with financial, health, consumer and data-protection regulators.

 

Companies should watch for definitions of covered systems, risk tiers, documentation duties, incident reporting, testing and remedies for affected users. The distribution of responsibility across model makers, application providers and deployers will determine whether the framework reaches only frontier developers or the much larger market of businesses using AI.

 

The next verifiable milestone is the consultation paper itself, due within one month of the October 8 announcement. Its publication should reveal whether India is proposing binding duties, voluntary codes or a hybrid model, and it will open the first meaningful test of how the country balances innovation, accountability and public protection.