Date: 16 February 2026
Venue: India AI Impact Summit, New Delhi

This dialogue examined what "safe AI" means in education, especially in the context of the Global South. Speakers included Srinivasan Parthasarathy (Professor of Computer Science and Engineering, The Ohio State University), Anil Ananthaswamy (Professor of Practice, IIT Madras), Swati Vasudevan (Country Director, Khan Academy India), Shaveta Sharma-Kukreja (CEO and Managing Director, Central Square Foundation), and Sunil Wadhwani (CEO and Co-Founder, iGate Corp). The session was moderated by Krishnan Narayanan (Co-Founder and President, itihaasa Research & Digital).

The panel discussed both safe-by-design features such as privacy protections and pedagogical guardrails, and safe-by-deployment approaches that rely on human supervision. A strong consensus emerged that teacher-directed, human-supervised AI is the safest and most practical model for education today, with student dependency and reduced critical thinking emerging as a major current risk.

The session highlighted that safety in the Global South must account for realities such as shared devices, low parental literacy, and severe educational gaps. Rather than waiting for perfect systems, the group argued for deploying AI responsibly to address urgent needs like foundational literacy, while ensuring that safety measures fit the local context. The broader goal is not just AI access, but AI fluency—helping teachers and students use AI with understanding and agency, in close integration with government systems and Digital Public Infrastructures.