This post is detailing our experience attending the AI Impact Summit and its associated side events in Delhi, February 2026. We are both unfamiliar with the policy and governance domain. This is just an honest reaction attending these events, maybe there are 2nd order effects we are not privy to and we welcome feedback.
Two events ran in parallel: the main AI Impact Summit and a separate, more curated side event called AI Safety Connect (AISC), co-hosted with the International Association for Safe and Ethical AI (IASEAI). We attended primarily governance and policy tracks at the main summit. On the second day, we also organized an independent AI safety mixer. This post covers what we observed across all three events, what worked, what didn't, and where we hope things would go in the future.
The Main Summit
What Worked
The summit's admission policy was open and inclusive. Late applications were accepted. Entry was free. Nobody was turned away. For an event of this scale that relates to AI, it really matters that this should not be a conversation restricted to a small group, and India has an opportunity to lead on broad, inclusive participation.
The summit was also useful as a reading of the room. By attending governance and policy sessions, we got insight into how different stakeholders, government officials, nonprofits, educators are orienting towards AI. We could hear what is salient to them: concerns about automation and job displacement, bias in AI systems, ownership of compute. Getting this kind of signal on world models and priorities across sectors was valuable.
I enjoyed the “Whose Language, Whose Model? Public-Interest Multilingual LLMs” by Aliya Bhatia, Center for Democracy & Technology Dhanaraj Thakur, Multiracial Democracy Project because it was a workshop format, this involved breaking up the room into rows of two and we all had discussions among ourselves. This format was much better than the panel version where we just got to listen and there