AI to empower not threaten: Meta AI scientist calls for new model architectures at IIT Delhi

NEW DELHI: Yann LeCun, Meta's Chief AI Scientist participated in a panel discussion at Indian Institute of Technology (IIT) Delhi on October 24, 2024. In the session titled "From Neural Mimics to Smart Assistants - A Journey into AI's Next Frontiers," LeCun outlined his vision for the future of AI and its integration with human capabilities.

He highlighted the necessity for a reevaluation of AI architectures, the benefits of open-source development, and India's distinctive potential within the AI sector.

The session ‘Be Inspired at IIT Delhi with Yann LeCun’ was opened by Dr Preeti Banzal, Scientist G & Advisor to the Principal Scientific Advisor, Government of India, who emphasised the Indian government's initiatives towards responsible AI.

Current AI frameworks not enough for physical world

During the discussion, Yann LeCun challenged existing methods of AI development, advocating for new architectural approaches that go beyond large language models (LLMs). He argued that the current AI frameworks are inadequate for achieving genuine human-like intelligence.

"We are not going to reach that stage by using the current paradigm and just making it bigger; we need essentially new architectures like objective-driven architecture”, Yann said. This approach focuses on creating AI systems designed to understand and interact with the physical world, enabling them to reason through new and unfamiliar situations effectively.

Also read IIT Delhi: 460 faculty posts vacant, but only 1-2% represent SC, ST, RTI data reveals

Yann highlighted that the future of AI depends on collaboration and innovation. With new architectures, open-source contributions, and India's diverse talent, AI can become a powerful tool for societal growth, empowering individuals and solving real-world challenges, he added.