AWS introducing AI trophy generator
AWS
The influence of generative artificial intelligence will be felt throughout this weekend’s Formula 1 Grand Prix du Canada. AWS, which is the title sponsor of the race, is introducing gen AI in three areas: a Statbot to fuel data-driven broadcast storytelling, a Root Cause Analysis to help F1 officials diagnose underlying operations problems and even an AI-generated winner’s trophy.
“We’re now starting to see the adoption of the technology in enterprise and in sports production,” said Neil Ralph, AWS Principal Sports Industry Specialist. “The examples we’re bringing to bear this weekend are examples of where that maturity is in terms of using the technology, effectively, to augment the human capability.”
The headline-grabbing, fan-facing activation is the trophy design. The AWS Generative AI Innovation Center helped craft ideas using the Titan Image Generator, with the AWS team seeding it relevant examples and experimenting with a series of prompts to harken back to its signature impact on the sport. (The full prompt can be found in this AWS post.)
AWS has been collaborating with F1 since 2018, including a series of computational fluid dynamics simulations that helped alter the air flow out of the back of the car — essentially, the goal was to change the aerodynamic wake in a way that would enable more passing in the sport. Ralph reported that, since the change in 2022, F1 has had 30% more overtaking.
“One of the main influences of the trophy design is this aerodynamic wake profile,” Ralph said, referring to the air “that’s pushed up and beyond the car that’s following. That vortex of air that flows over a Formula One car is a significant influence in the trophy design.”
AWS
A few other uniquely Canadian details — the maple leaf and St. Lawrence River — were invoked in the creation of a 3D model that a traditional silversmith then used to forge the actual trophy. Fans, meanwhile, are invited to use a publicly accessible model, the PartyRock Trophy Generator, to create their own versions.
The AWS Statbot leverages gen AI to enable natural language prompts to query F1’s database, which includes all of its racing statistics since 1950. That will be available to the centralized broadcast production group in order to aid analysts in finding factoids and precedents to share swiftly.
“You want the answer to that question as quickly as possible, so you can get it on the broadcast to be editorially relevant,” Ralph said. “With Statbot, what we’ve done is augment that human capability and have the opportunity for using natural language prompts get answers back from the historic data repository far quicker than ever could be done by humans.”
This serves as an extension of the AWS-powered F1 Insights program, which has created 23 metrics over the past half-dozen years. For now, Statbot acts as a responsive query, but Ralph indicated a future iteration on the product roadmap is for the tool to be proactive in identifying and suggesting information. He also said AWS would love to create a version in which fans can ask their own questions.
The third implementation of gen AI is to aid F1’s operational services. Whether it’s related to the broadcast production or in race governance capabilities, Formula 1 officials sometimes need to race to fix problems quickly during a race weekend. That sometimes means treating symptoms rather than eradicating the core issue, which is where the Root Cause Analysis solution will enable the circuit to pore through extensive log files far more efficiently than a person could to diagnose the underlying error.
