Summary
IT Minister Ashwini Vaishnaw said the government will release a consultation paper on AI regulation within a month, declaring that the time has come for such a framework. It will focus on AI safety, deepfakes, a human-first approach, skilling and inclusion and the psychological impact of AI.
The Minister said AI cannot be regulated by law alone and called for a techno-legal solution, with industry bearing the bulk of the responsibility. A key area is expected to be agentic AI: how much autonomy such systems should have and whether they may retain and reuse data.
The announcement reverses the government's earlier view that existing laws were sufficient, after piecemeal steps such as an April proposal requiring platforms to label AI-generated content continuously.
WHY IT MATTERS FOR UPSC & STATE PCS
The Union IT Minister has announced a consultation paper on AI regulation within a month, signalling a shift from relying on existing laws to a dedicated AI framework. Deepfakes and agentic AI are expected to be central.
Standard News
A Deepfake Is a Bad Output. An Agent Is a Bad Action.
Here's what is actually happening. Until now, India regulated AI mainly by looking at what it produces: a fake video, a cloned voice, a misleading image. The coming consultation paper will have to regulate something harder: what AI does when nobody is watching each step.
The policy shift in one line For years, the government's position was that India did not need an AI law and that the IT Act, 2000 and its rules could handle the problems.
On Thursday, IT Minister Ashwini Vaishnaw reversed that: the time for AI regulation has come. A consultation paper will follow within a month, focusing on safety, deepfakes, a human-first approach, skilling and the psychological impact of AI.
The method matters as much as the decision. The Minister called for a techno-legal solution, meaning law plus technical safeguards built into the systems themselves and said the bulk of the responsibility must lie with industry.
That is an admission that a statute alone cannot keep pace with the technology.
Why labels work for deepfakes A deepfake is an output.
Someone generates a fake image, video or voice and it spreads. The harm lies in people believing it. That makes deepfakes regulable at the point of display. Label the content as synthetic, require platforms to show the label continuously, as proposed in April and take down flagged material quickly.
The tools are imperfect, but they fit the problem, because they act on the same thing that causes the harm: the content people see.
Why labels fail for agentic AI An agentic AI system does not just generate text.
It acts. It can search, decide, click, pay, send and store, across many steps, without a human approving each one. Here the harm is not in anything displayed. It lies in an action taken without permission or in data collected and retained and reused in ways the user never expected.
A label on the screen solves neither. That is why the government is reportedly weighing two specific questions: how much autonomy agents should have and whether they may keep and reuse data. A useful comparison is a power of attorney: you authorise someone to act for you within defined limits.
The comparison breaks down at one point. A human attorney can be held legally responsible; an AI agent cannot. So rules for agents have to fix responsibility on the developer or deployer through permission limits, action logs, human checkpoints for high-stakes steps and limits on data retention, which fits naturally with India's Digital Personal Data Protection Act, 2023.
Where
India stands The European Union has already adopted a risk-based AI law. India is choosing a lighter, techno-legal route that leans on industry responsibility. It is late to dedicated regulation but early enough to design it for the agentic era rather than retrofit it. For GS3, the takeaway is simple: regulate outputs with transparency, regulate actions with permissions and accountability.
Quick Facts
Key numbers & takeaways — revise these first
-
Announcement: consultation paper on AI regulation, to be released within a month.
-
Announced by: Union Minister of Electronics and IT, Ashwini Vaishnaw, at the launch of the World Bank's World Development Report 2026 in New Delhi.
-
Key focus areas: AI safety, deepfakes, a human-first approach, skilling and inclusion and psychological impact.
-
Approach: techno-legal, combining law with technical safeguards; bulk of responsibility on industry.
-
Policy shift: the government earlier said existing laws were enough to deal with AI.
-
Earlier step: an April proposal requiring platforms such as YouTube, Instagram and X to display labels continuously on AI-generated content.
-
Options explored earlier: a consent-based framework for synthetic content, curbs on agentic AI autonomy and regulatory sandboxes for high-risk uses.
-
Agentic AI: autonomous AI systems that go beyond generating text and can carry out tasks.
-
Existing laws: Information Technology Act, 2000 and the IT (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021.
-
Nodal ministry: Ministry of Electronics and Information Technology (MeitY).
Connect the dots for your UPSC preparation.
Standard news covers the event. Log in to read our comprehensive analysis and uncover the hidden constitutional, structural, and ethical dimensions of this topic:
A clear breakdown of why deepfakes are an output problem and agentic AI is an action problem and why each needs a different regulatory tool.
Four structural causes behind India's reversal from "existing laws suffice" to a dedicated AI framework.
What a techno-legal regime for agents could include in practice: permission scopes, action logs, human checkpoints and data-retention limits.
Where India's industry-led approach sits against the EU's risk-based model and the risks of leaving the bulk of responsibility with industry.
Included in this analysis
Join thousands of aspirants analyzing the news deeply.
Unlock Premium — Rs.699 AnnuallyDon't have an account? Sign up for free