Karnataka Proposes AI-Driven Social Media Safety Bill

Karnataka has drafted the Karnataka Responsible Social Media & Digital Safety Bill, 2026, prepared by the Karnataka State Policy and Planning Commission and submitted to Chief Minister Siddaramaiah. The draft creates a state-level, AI-focused regulatory framework that mandates labeling of AI-generated content and deepfakes, sets enforceable takedown windows of 24-48 hours, and establishes the Karnataka Digital Safety & Social Media Regulatory Authority to monitor compliance and adjudicate grievances. The bill combines rapid, technology-assisted moderation tools with a public digital literacy program and phased implementation. Its timeline targets legal vetting and a monsoon-session introduction in June-July 2026. For practitioners, the bill signals tighter, time-bound moderation requirements, a push for synthetic-content detection and provenance metadata, and potential state-level enforcement that could complicate compliance for platforms operating across India.
What happened
Karnataka released the draft Karnataka Responsible Social Media & Digital Safety Bill, 2026, prepared by the Karnataka State Policy and Planning Commission and submitted to Chief Minister Siddaramaiah. The draft mandates labeling of AI-generated content and deepfakes, imposes binding removal windows of 24-48 hours, and proposes the Karnataka Digital Safety & Social Media Regulatory Authority to enforce compliance and handle grievances. The bill phases implementation with initial awareness and institution setup, followed by tech integration and full enforcement, targeting legal vetting and introduction in the June-July 2026 monsoon session.
Technical details
The draft centers on operationalizing AI tools and timelines rather than high-level obligations. Key elements include:
- •Mandatory labeling and precise legal definitions for synthetic content and deepfakes
- •Enforceable takedown windows of 24-48 hours for harmful content, shifting from advisory central guidelines to state-level deadlines
- •A state regulatory authority responsible for monitoring, real-time dashboards, and compliance actions
- •User rights: time-bound grievance redressal, reporting mechanisms, and protections against harassment and misinformation
- •A digital awareness and media literacy program aimed at prevention and youth mental-health concerns
Platforms will need to combine automated detection models with expedited human review, provenance metadata, and faster incident workflows. The draft references tech tools such as fake-news detection and deepfake tracking; it also contemplates penalties for misuse, though the bill text will need legal vetting to detail audit, reporting, and sanction mechanisms.
Context and significance
This is the first Indian state-level proposal that explicitly ties digital-safety rules to AI-specific obligations. It departs from the federal Information Technology Act, 2000 and IT Rules, 2021, which emphasize due diligence and notice-and-takedown without strict short windows. For platform engineers and policy teams, the bill signals rising demand for low-latency moderation pipelines, localized language models, robust provenance tagging, and operational transparency. It also raises jurisdictional questions: overlapping state and central rules could force platforms to reconcile multiple compliance regimes, potentially increasing legal and engineering complexity.
What to watch
Track the legal vetting outcomes, the authority's precise enforcement powers, and whether the center responds with harmonized federal guidance. Operationally, watch for guidance on metrics, auditability, appeals, and resource expectations for smaller platforms.
Scoring Rationale
State-level, AI-specific regulation is notable because it creates a new compliance vector and could set a precedent, but it is not a national or global paradigm shift. The proposal is actionable for engineers and policy teams, hence a mid-high importance score.
Practice interview problems based on real data
1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problemsStep-by-step roadmaps from zero to job-ready — curated courses, salary data, and the exact learning order that gets you hired.



