The “FerrAI” isn’t just an idea but a reality: Ferrari has been utilizing it for six years now. It’s not just for show, as demonstrated by the recent progress of the SF-24.
It didn’t take much to replace a Ferrari driver and proudly showcase my Boston Terrier seated in the SF-24 cockpit. It’s a silly application of artificial intelligence, of course, and we’re no longer immune to such foolishness. However, AI, also used to spread fake news, is a profoundly serious matter, now integrated into Formula One for six years, back when nobody talked about it. It’s not just an accessory but increasingly strategic and crucial as it advances. Rapidly evolving over the past five years, it reached maturity two to three years ago. In Maranello, it’s making a difference and is the foundation of the recent decisive progress of the Scuderia. There are no external suppliers or consultants; everything is developed in-house, enhancing the shroud of secrecy surrounding these projects. This is the true frontier on which teams operate (still) freely, without regulatory constraints.
How many people are involved in artificial intelligence at Ferrari, and who leads this working group?
“AI within Maranello has a name and a surname – Marco Adurno – and there are four minds dedicated to it, no more. To be applied, the conclusions are then extended to the entire Performance Group, around twenty engineers.”
When did it start?
“In 2020, with the hiring of Adurno from Red Bull, a forty-five-year-old Milanese. A highly valuable addition. Linked in 2019, Adurno managed to join in September, and the project was initiated at the beginning of 2020.”
How has AI progressed at Ferrari?
“Two years were spent on experimentation and refinement, so the plan began to work in 2022, with a significant improvement in correlation and related work methodologies from mid-2023. The fruits of this labor have been apparent for a couple of years, but the first major performance leap attributable to AI is this season, evident with the SF-24. We can define this car as the first competitive Ferrari (also) thanks to AI.”
What application does AI find?
“It’s used to eliminate discrepancies between wind tunnel data and track performance. We’re talking about the famous correlation, which is notoriously difficult to achieve: with AI, the latest data from all simulations and the track are inputted, and using those, artificial intelligence identifies deviations and suggests corrective actions. Technicians adjust the values so that AI – to put it simply – knows more than before, thus leading to further improvement in simulations. This allows the Scuderia, once on the track, to perform immediately as if they had physically tested, not just virtually. Charles Leclerc demonstrated this clearly by turning around his weekend and engaging in a fierce battle with Max Verstappen and Red Bull, despite having to forego the free practice sessions entirely. A situation never seen before.”
In a nutshell, how can this use of AI be explained?
“AI discovers how to better operate the simulator, even though it has never seen it operate in that way before.”
Has it ever been wrong?
“No, since its development, it has never pointed in the wrong direction in perfecting correlation.”
Who’s ahead?
“Ferrari and Red Bull, not necessarily in that order. Mercedes is trying every way possible…”
Leave a Reply