After criticism, India's Ministry of Electronics and IT revises its AI advisory to drop plans to require government approval before new AI models are launched
Context & Ripple Effects
MeitY had issued a March advisory seeking explicit government permission before developers tested or deployed AI models or algorithms. The reversal retreats from that proposed pre-clearance regime after criticism, restoring a less restrictive near-term posture than the ministry's earlier move suggested.
The change matters because it reopens the gap between India's stated support for AI growth and the use of platform-style controls in adjacent digital policy, including its misleading-content requirements for social platforms.
First-order effects
- AI developers no longer face the proposed requirement to obtain government approval before launching new models under the revised advisory, removing an immediate compliance and timing hurdle.
- MeitY preserves room to issue AI guidance while stepping back from a mechanism that would have put model releases behind a government gate.
Second-order effects
- The reversal shifts the policy debate from pre-launch licensing toward what narrower obligations, if any, should govern developers after deployment; companies will still need to monitor advisory language closely.
- Criticism has demonstrated that a broad approval mandate can be politically and commercially difficult to sustain, raising the bar for future interventionist AI rules.
Third-order effects
- If this pattern holds, India's AI regime may develop through iterative advisories and targeted safeguards rather than a single blanket approval system, leaving developers with flexibility but less regulatory certainty.
- The episode is a test of whether India can align AI oversight with its earlier position that AI is strategically important and should not face growth-stifling regulation, as expressed in the ministry's 2023 stance on AI growth.
The trend: India is moving toward contested, state-mediated AI governance in which ambitions to foster domestic AI capability must be balanced against demands for tighter controls.