GenerativeAI is beginning to be pervasive, although we must not forget PredictiveAI. The big question is Sovereign AI: where India should invest its limited generativeAI funds: in foundational models, or in the layers above and below? The layer above: use open source models, focus on applications in domains such as agriculture, medicine, defense, jurisprudence and governance? The layer below: data centers hosting multiple models? Can't ignore concerns about job loss, water/energy demand from data centers, biases from training/post training data that skew model output, hallucinations; loss of cultural and national sovereignty, based on western or chinese values or politically charged (left or right) perspectives baked into the models".
Your question goes straight to the speakers.
