AI: The good, the bad, and the life saving

Leslie's research is about "Responsible AI", and this show is an attempt to explore it with the public. Tech bros seem determined to boast that their AIs are so powerful that they could destroy the world - just what are they compensating for? The MacGuffin that they are chasing - artificial general intelligence or AGI - made Rishi Sunak set up a Global AI Safety summit at Bletchley Park, calling all the world leaders to Milton Keynes in what can only be described as an austerity remake of Avengers Assemble. In this talk he'll examine the AI that we want vs the AI that is being sold to the world.

AI has recently shown how powerful it can be, but what about the most important milestone of intelligence: Mario Kart. Reinforcement Learning allows AI to repeatedly try a task and learn from their mistakes. This talk will present Tyler on his YouTube channel where he has built world-leading AI to play Mario Kart. This talk will explain how AI can go about learning these complex games that humans love - or in the case of the infamous Rainbow Road, hate!

Destructive wildfires cause widespread devastation across the planet. Detecting these and preventing them from spreading is a huge challenge and one which requires bold new solutions. The Flare-X team pulls together expertise from across the world to tackle this problem as part of the $11M X-Prize wildfire challenge. This talk will explain a multilevel approach currently in development to locate and extinguish incipient-stage wildfires with a fully autonomous swarm of aerial drones, as well as laying out the current and future demonstrations and deployments of this approach.