India drops plan to require approval for AI model launches

India has backtracked on a recent AI advisory that required government approval before launching or deploying AI models, amid criticism from entrepreneurs and investors. The Ministry of Electronics and IT unveiled an updated advisory that no longer mandates approval but urges firms to label unreliable AI models. The move follows backlash from industry figures, including Martin Casado of Andreessen Horowitz. India’s previous hands-off AI regulation stance has shifted, with the advisory emphasizing compliance without legal binding. The guidelines stress avoiding unlawful content sharing and safeguarding against bias and election interference, while encouraging the explicit notification of AI-generated output unreliability. Maintaining a focus on combating deepfakes and misinformation, the ministry advises using unique identifiers rather than identifying message origins.