VARANASI, June 15 -- Prof Ashutosh Sharma, president of the Indian National Science Academy (INSA), asserted that "man is not equal to machine," highlighting the critical need for an "AI for India that knows what it does not know." He was speaking on the second day of the 7th National Convention of Vijnana Bharti, which featured a curated panel discussion on "AI and Ethics." Led by Prof Sharma, the session was moderated by Dr Vivek Kumar Singh, senior advisor at NITI Aayog. The panel featured experts, including Prof Mayank Vatsa of IIT Jodhpur; Deepak Maheshwari, senior policy advisor at CSEP; and Dr Maneesh Singh, VP of Machine Learning at Bharat Gen, IIT Bombay. Prof Sharma observed that while technological advancements are accelerating, societal and human adaptation continues to lag. He advocated for ethical AI characterised by human oversight, transparency, accountability, fairness and privacy. He emphasised the necessity for an AI framework tailored to the Indian context, one that is inclusive, multilingual and socially beneficial, particularly for communities currently underserved by conventional systems. Ultimately, he called for a deeper understanding of human consciousness alongside AI, guided by both ethical principles and societal responsibility. Dr Maneesh Singh argued against simply adopting global AI models. Instead, he proposed developing a transparent, sustainable and citizen-centric ecosystem that respects data rights and recognises the limitations of machine intelligence. He stressed that the objective should be to cultivate "AI for India" rather than merely importing foreign solutions. Prof Mayank Vatsa discussed the growing role of AI in healthcare and medical science. He cautioned, however, that AI should serve as a supportive tool rather than a replacement for professional medical advice. Dr Madan Mohan Tripathi, director general of the National Institute of Electronics and Information Technology (NIELIT), noted the rapid proliferation of AI and the inevitable need for regulation. Comparing AI to technologies like GPS, he suggested that risk-based frameworks ranging from outright restrictions to controlled usage are essential. He added that since most global models, including GPT, are trained on non-Indian datasets, domestic efforts to develop localised models are critical....