Can our AI systems be far less energy-hungry without sacrificing performance? In his talk, Roeland Nusselder, a computer scientist and the CEO of Plumerai, explores how the growing scale of AI models, such as those used in machine learning and natural language processing, are becoming ever more resource-intensive. He goes on to show how the rapid development of AI technologies could potentially overwhelm our current energy infrastructure, unless we make significant innovations to reduce their energy consumption. To counter this trend, Mr Nusselder introduces the concept of "tiny AI", or AI systems that are optimized to be much smaller, more efficient, and less energy-hungry without sacrificing high performance. What exactly does tiny AI involve?