Formula 1 (F1) is as much a battle between the world’s best engineers as it is the world’s best drivers. With pit stops being taken in under 2 seconds and drivers hitting speeds as high as 370kmh and risking their lives, F1 needs a technology as fast as its racers. Unsurprisingly, it has positioned itself as the most technologically advanced sport, actively pioneering innovations that are even adopted in other industries.

Few things are truly unpredictable. If it can be quantified it can be predicted. The world of Formula 1 (F1) is rich in quantifiable features, making it ideal for prediction and with 120 sensors on each car generating 1500 data points per second, the sport is highly compatible with Machine Learning (ML) which can handle large complex datasets. Thus, as the world’s most data-rich sport, it has naturally become the fastest growing sport for ML. 

F1’s ML models are used to evaluate 65 years of historical race data, make forecasts, and display insights into the split-second decisions and strategies of teams and drivers during races. As such, success in F1 is now increasingly determined by the use of data to calculate, analyze, and design optimal strategies in an iterative process to continually improve and try to outperform the competition.

 

AI: The Next Step for F1?

As team boss, Christian Horner explains: “AI and ML are big categories that are emerging. Both areas, with the amount of data that we generate, the way that we simulate, etc, are going to play a key role in our decision making as track time becomes ever less”. Horner’s Red Bull has identified the potential of ML to advance in-race strategic decision-making, deciding to team up with technology giant Oracle. “This year we have had three days of testing, and no other sport would have such a small amount of practice. So the way that we analyze the data is crucial for us, and I think this is where this partnership is going to pay absolute dividends.”

AI can also be used to enhance the fan experience. Whether it be by determining what kinds of promotional content individual fans respond to or by using millions of data points from racing metrics to engage fans by illustrating the nuances of split-second decision-making through F1 Insights, a newly added component of race information displayed enabled by Amazon Web Services (AWS). 

Even off the racetrack, AI could help to propel the sport to new heights. AI can analyze individual components during manufacturing to identify weaknesses that can be improved in the build design, enabling lower-budget teams to close the gap with leaders. For instance, F1 has combined with AWS, running aerodynamic simulations to develop a car that will be introduced in 2022 that reduces downforce loss from 50% to 15% , giving chasing drivers greater chances of overtaking and encouraging more wheel-to-wheel action. 

Addressing F1’s Jeopardy Problem with Cloud computing:

F1 has a jeopardy problem. For drivers, at constant risk of serious injury and even death, there is too much peril. For viewers, on the other hand, there is not enough. With only three drivers (Lewis Hamilton, Nico Rosberg, and Sebastian Vettel) and two constructors (Mercedes and Red Bull) having won championships since 2010, it has become one of the world’s most predictable sports. As a result, F1 has turned its attention to cloud computing to address both sides of the jeopardy problem.

Through the AWS cloud and using 1150 computer cores to simulate more than 550 million variables affecting downforce, F1 has modelled a car for the 2022 season that will make close-quarters driving and overtaking easier by reducing downforce loss from driving in the turbulent air of the car ahead. This will make fighting for positions safer for drivers whilst equipping them to engage in more ‘extreme’ seemingly dangerous maneuvers that will excite viewers. 

 

Ethical concerns of AI in F1- The Downsides of Datafication:

Younger fans with increasingly diverted attention spans will no longer be appealed by F1’s traditional 2-hour linear feed. To ensure the sustainability of the sport, F1 has to make use of its vast pool of data to tell stories in a way that this generation will consume. Datafying the sport, however, could lessen the excitement of viewers and muddle the competitive advantages of some F1 teams.

Viewers are now presented with a much wider range of race insights. Battle Forecast, for instance, uses track history and projected driver paces to predict how many laps before the chasing car is within ‘striking distance’ of the car in front. Does removing the uncertainty and suspense not make the chase less captivating? Last year alone, six statistics—Exit Speed, Predicted Pit Stop Strategy, Pit Window, Battle Forecast, Pit Strategy Battle, and Tyre Performance–were added to F1 broadcasts. 

The availability of this much data over-informs viewers, allowing them to know or reasonably estimate race outcomes beforehand which dilutes the jeopardy component that keeps them engrossed. As Smedley explains “ what you don’t want to do is tell the story before the story has emerged. Sport is all about jeopardy, and you don’t want to take that jeopardy away by us being super data analysts and data scientists.” The AI “is to engage fans, this is to get fans to sit on the edge of their seat, not to do the opposite,”. Now even if the information is kept from consumers to maintain F1’s jeopardy, does this then become a question of determinism vs free will if computer scientists can accurately know who the race winner will be before the checkered flag? Is there value lost in the outcome of a race if you are aware of others knowing the result? Like how you might prefer to watch an event live rather than pre-recorded regardless of your awareness of the result.

While the datafication of F1 could diminish the entertainment value of the sport on the consumer side, it also has drawbacks for the racing itself. What if a team’s competitive advantage is inadvertently revealed when the insights graphics are broadcasted live? Or what if a driver or car receives consistently low scores in the data insights – will that put off sponsors?

Conclusion:

In every race, teams are faced with the challenge of formulating strategies to maximize their chances of success. When should their drivers pit? What car settings would work best for the track? Traditionally, these decisions are made by race engineers who interpret real-time data and conclude the best course of action. However, when massive sums of money are on the line and when a press of a button can be the difference between standing on the podium and watching another driver lift the trophy from the team garage, there is no space for error, which is why F1 seems so well-poised for the integration of AI. 

Although AI has the potential to revolutionize F1 by optimizing race strategies, fan experience, and vehicle designs, making the sport more competitive and fun to watch, it can have an equally detrimental impact by reducing viewer excitement and removing teams’ competitive advantages. This balance has to be carefully considered in light of how promising albeit potentially harmful AI is for F1 and whatever decision is made moving forwards will surely create plentiful opportunities for the sport to reach new heights at the pinnacle of motorsport. After all, we have to understand how the track is laid-out before deciding when to pit and what tyres to use. 

LEARN MORE ABOUT AI

Operationalising AI Ethics in 2021
Discussing the ethical implications of AI and how to regulate them has been the central topic of our conference for the past 4 years, but we have never really defined what we mean by ethics! In our...
Governance and Accountability in AI
Today, we look back on our last panel series = governance and AI. We hosted Katherine Mayes, the Head of Data Analytics, AI and digital ID at TechUK, Ellis Parry, the Information Commissioner’s...
What is AI for Good?
AI is often discussed in the context of incorporating ethics into automation or data analysis. This is different from AI for Good. At the AIBE Summit 2020, Manoj Saxena spoke about AI Global, which...
Checklists and frameworks for ethical AI design in business
While the top ethical issues in AI have been identified and are fairly well-known, we often do not know what this means on a technical level. Our ability to think through the consequences needs to...
AI – A weapon against the climate crisis?
There is no use denying it, climate crisis is real. The Anthropocene, said to be a new geological epoch in which man is the main actor on all natural changes, the period we are living in is scary. ...