Union electronics and information technology minister Ashwini Vaishnaw said on 8 October that MeitY will release a regulation-focused artificial intelligence consultation paper within about a month, framing safety, deepfake harms and skilling as core pillars while insisting that industry—not statute alone—must carry primary compliance duty.
Techno-legal framing
Speaking at the launch of the World Bank’s World Development Report 2026 on AI, Vaishnaw argued that pure legislation cannot keep pace with model releases. He repeated the government’s preference for techno-legal tools that combine IT Act-era rules with sector standards, echoing the Artificial Intelligence Governance and Economic Group’s cross-ministry coordination mandate.
The minister listed cybersecurity, psychological harm and synthetic media as risks requiring documented mitigations before models reach consumers at scale. Stakeholder comments on the upcoming paper will likely feed into MeitY’s existing IT Rules 2021 enforcement practice rather than a standalone AI Act in the next quarter.
Compute and indigenous hardware
Vaishnaw also outlined a five-year procurement plan for 10,000 graphics processing units under the IndiaAI Mission, with an initial tender covering 5,000 units to give cloud providers demand visibility. Continuous empanelment every quarter lets new bidders supply compute to researchers and startups at subsidised rates.
He teased a three-to-four-year roadmap for indigenous GPUs tuned for inference workloads, a politically sensitive goal as import dependence on Nvidia and AMD remains high. Details are expected in coming months alongside empanelment of AI models for government departmental use.
What builders should do now
Indian startups hosting chatbots or voice agents should document training data provenance and incident response contacts before regulators ask. Enterprises deploying copilots for customer service need watermarking and escalation paths when models hallucinate policy answers.
Global vendors selling into India will watch whether the consultation paper aligns with the EU AI Act’s risk tiers or stays lighter, preserving MeitY’s pro-innovation rhetoric. Either way, Vaishnaw’s emphasis on industry responsibility signals more audit requests, not a pause in public-sector AI pilots.
Limits
A consultation paper is not law. Court challenges to deepfake misuse will still run through existing penal provisions and IT Rules takedown orders. The desk will revisit when MeitY publishes the draft text and comment deadline.
Enforcement reality
MeitY already uses IT Rules takedown orders for deepfake clips during election seasons; the consultation paper likely codifies notice-and-action timelines rather than inventing a new regulator. State governments running chatbots for grievance redress will need model cards describing training cut-off dates and refusal policies.
Startups selling synthetic voice for call centres should expect questions on consent logs and biometric data storage, especially after RBI’s parallel push on digital lending KYC. The GPU tender’s five-year horizon may favour incumbents who can finance datacentre builds in Chennai and Hyderabad cloud regions.
Legal teams at broadcasters want clarity on liability when newsrooms air user-generated clips that later prove synthetic; the consultation paper is expected to address intermediary safe harbour alongside creator duties.