Why AI is Power-Hungry—Lessons from the Brain and How We Can Fix It
Artificial intelligence has made remarkable strides, yet it remains fundamentally distinct from biological intelligence—particularly in energy efficiency. While the human brain operates at approximately 20 watts, modern AI systems consume vastly more power. Rather than scaling AI with ever-increasing energy demands, we should explore alternative computational frameworks inspired by biological intelligence. This talk will highlight the fundamental energy challenge AI faces and explore how rethinking learning algorithms and AI hardware can lead to more efficient systems.
Date:
20 February 2025, 15:00
Venue:
Sherrington Building, off Parks Road OX1 3PT
Venue Details:
Florence Buchanan Lecture Theatre
Speaker:
Dr Luke Taylor (University of Oxford)
Organising department:
Medical Sciences Division
Part of:
Oxford Neurotheory Forum
Booking required?:
Not required
Audience:
Members of the University only
Editors:
Ian Cone,
Jascha Achterberg
