‘Time has come for AI regulation’, says Vaishnaw; government to release consultation paper

Government to Consult on AI Regulation

IT Minister Ashwini Vaishnaw said on Thursday that the Centre will publish a consultation paper on artificial intelligence within a month, with the stated aim of regulating the technology. The announcement marks a shift from the government’s earlier position that existing laws were sufficient to address AI-related challenges. The consultation will centre on AI safety, skilling and inclusive adoption, with Vaishnaw stating that humanity must remain central to the approach and that the bulk of responsibility should rest with industry.

Deepfakes and Agentic AI in Focus

The paper is expected to examine risks from deepfakes, whose quality Vaishnaw said is improving daily and misleading users, as well as the psychological impact of AI on users. A key area is agentic AI — autonomous systems that plan multi-step actions, use external tools and adjust strategies with minimal human oversight. Questions under discussion include how much autonomy such agents should have, and whether they may retain and reuse data. Liability of AI platforms and developers for model-generated content is also under examination, since safe harbour immunity for user-generated content on social media does not translate cleanly to AI models trained on internet data.

What Businesses Should Watch

The government has so far proceeded through piecemeal measures, including an April proposal requiring platforms such as YouTube, Instagram and X to continuously label AI-generated content. A broader framework follows concerns over high-quality objectionable deepfakes, agentic AI and cybersecurity risks. Reports from July indicated the IT Ministry was exploring a consent-based framework for synthetically generated content, curbs on agentic AI autonomy and regulatory sandboxes for high-risk applications. Businesses deploying AI, particularly in content, financial services and consumer-facing roles, may face new labelling, consent, data retention and liability obligation…