In last week’s newsletter, we explored the rapid advancements in AI’s mathematical capabilities. This week, that progress has been dramatically underscored as a Google DeepMind system achieved silver at the International Mathematical Olympiad (IMO). By solving four out of six problems, including the competition’s hardest question, the system demonstrated its ability to tackle challenges at a level comparable to the world’s brightest mathematicians. This achievement is particularly significant in terms of the mix of techniques used, combining LLMs, reinforcement learning, and formal mathematical proofs in a novel way. This synergy of different approaches demonstrates how complex problems can be tackled by leveraging inventive architectures. Its also interesting to see Gemini playing a key role in the solution, translating natural language problem statements into formal mathematical statements to help other AI components train themselves. More details will no doubt emerge, but this looks like another notable breakthrough by the DeepMind team.
Meanwhile AI’s influence isn’t confined to the realm of the abstract. As the sporting Olympiad gets underway in Paris today, AI is set to transform how we experience, participate in, and safeguard major sporting events. The International Olympic Committee (IOC) has identified over 180 potential use cases for AI in these games. Google is partnering with NBC and Team USA and will leverage Gemini to enhance viewer experience. Commentators will use AI-powered tools to provide quick insights into sports, athletes, and rules, making the games more accessible and engaging for audiences worldwide. NBC’s “chief superfan commentator,” Leslie Jones, will use Gemini to explore and explain new sports in real-time, adding a fresh dimension to Olympic coverage.
AI’s role in enhancing viewer experiences and providing personalised content could revolutionise how media companies engage with their audiences. NBCUniversal’s “Your Daily Olympic Recap,” which offers AI-generated summaries of favourite events narrated by an AI version of legendary sports presenter Al Michaels, points to a future where content is tailored to individual preferences at scale. AI is also pivotal in bolstering security around Paris. Systems have been deployed to monitor crowds and detect suspicious activities. This sophisticated surveillance system is designed to pre-emptively identify potential threats, providing an additional layer of security beyond traditional methods. Sports analysis is also in scope for enhancement: technologies like Intel’s 3D athlete tracking (3DAT) analyses biomechanics, potentially levelling the playing field by helping coaches identify talent and optimise training regimens.
Takeaways: AI’s Olympic prowess is diverse. Progress in maths competitions presages more powerful reasoning and scientific uses. Use in sporting events shows versatility but critics will rightly worry about privacy infringement and the normalisation of surveillance. This tension between security and civil liberties exemplifies the broader challenges of AI adoption. The Paris Olympics will no doubt keep us on the edge of our seats for the next two weeks, but it will also serve as a significant test case for the integration of AI across various aspects of a large global event.