By now, if you’re a dedicated Formula One motor racing fan, you’ll have noticed the AWS logos popping up. They’re on the Ferrari cars, as part of a sponsorship deal, but they also appear as on-screen graphics, highlighting what’s happening on the track.
AWS is, of course, Amazon Web Services and it’s not the bit of Amazon that sells you books or on-demand videos. AWS is Amazon’s bespoke computing department, and it uses cloud-based technology to, essentially, create supercomputers in the sky – tying together the computing power of hundreds of individual machines to generate ultrarapid responses.
So far, AWS’s contribution to Formula One has been in helping to analyse what’s happening on track, and potentially to predict what’s going to happen next. AWS can analyse in real time whether one driver is lapping faster than another, and can predict when those two cars might be close together and we might actually see some racing.
Before that, AWS’s computing power was used by the sport’s rule-makers to help develop the aerodynamic rules first introduced in 2022. These were designed to help the cars run more closely together and to encourage more overtaking and on-track action.
Now, though, AWS is going one step further, and has collected and collated every last shred of data and stat from the earliest days of Formula One right up to today, in an effort to provide broadcasters and commentators with instant access to any relevant piece of information that they might need in the heat of live television.
Neil Ralph is the sports industry specialist for AWS and he told The Irish Times that: “The work we’ve been doing most recently with Formula One is on how to leverage our capabilities to make the broadcast team more effective. So, F1 has existed since the 1950s, and we have 70 years of historic data related to results, lap times, qualifying and so on. Previously we’d have built up a package of facts and figures that might be relevant to an upcoming race weekend, but if something came up during the race, we’d have to go back and query the database again, and that could take several minutes, during which time you might have lost the editorial relevance because the live situation has changed.”
Amazon wanted to find a way to speed this process up, and the answer is Statbot. The bot is an AI tool that allows all-out instantaneous answers to questions about F1 history, and can flash up the necessary info to commentary teams as the race is in progress.
“Today, through Statbot, we have the ability to orchestrate that large data set through natural language queries” says Ralph. “It allows broadcast teams to ask the questions that they need answers to, and it allows Amazon to put those queries against the historical data repository and get those answers back in literal seconds rather than minutes, and maintain editorial relevance and improve the broadcast product for the fan base.”
I remember peering into the back of the old broadcast booths at Silverstone once to see the late, great Murray Walker – the British F1 commentator par excellence – and his vast board of notes for the race, detailing useful stats and facts about each driver. Now, Amazon seems to have recreated Murray’s brain and board in digital form – albeit while Statbot can tell you which was the closest race finish in F1 history (the 1971 Italian Grand Prix, 0.010 sec between Peter Gethin and Ronnie Peterson, and no I didn’t have to look that up ...) it can’t tell you which F1 driver’s family owns the best ice cream parlour in Siena (it’s Alessandro Nannini). Stats are useful, but they don’t tell you the human story.
Then again, could Statbot eventually replace racing altogether? After all, if it knows everything about F1 history, and AWS is currently recording data from each race and each car and each lap, could Statbot not just accurately predict the outcome of each race and we could all save ourselves the bother?
Thankfully not, it seems. “I think the short answer is no” says Ralph. “Statbot is there to augment our assistance to broadcasters. Our in-race insights are developed with people who are recognised experts in the field, and we’re here to tell data-driven stories, stories that improve the experience of the global fan base of F1. And what we’re now shifting towards is listening more to the data, allowing the data to help identify something that might be interesting, maybe down in the mid-field of the race, where the cameras haven’t spotted it, and flag that up to the broadcast teams. For me, it’s about augmentation, the artificial intelligence is augmenting human intelligence, but it’s not about replacement.”
At least David Croft and Martin Brundle are safe in their employment for now. Instead of trying to replace the human element, AWS at least claims that what it’s trying to do is to make the work of the teams that tell the stories of F1 that bit easier. It’s also delving into other areas, such as problem-solving technical issues during live broadcasts so that everyone at home can enjoy a more seamless experience, and even encouraging fans to use generative AI to design new race winner’s trophies, with the prize of a trip to an F1 race in 2025 being offered.
AWS has even tried to solve the biggest conundrum in Formula One history – who was the fastest, ever? A few years back a team that included former Ferrari F1 chief mechanic Rob Smedley used AWS data to crunch the data on every driver’s qualifying laps – when the cars are at their lightest and fastest, generally – to see if it could work out who was actually the quickest driver of all time.
The answer? The late Ayrton Senna. That was perhaps not surprising. Even away from data-sets, the great Brazilian was generally considered by many to be the fastest qualifier of all time, and for many years – even posthumously – he held the record for the greatest number of pole positions, until he was eventually overhauled by Michael Schumacher, and more recently by Lewis Hamilton.
However, the data used by AWS was somewhat flawed, one might argue, especially as it only went back as far as 1983. Before that, AWS argued, timing wasn’t precise enough to come up with a definitive answer but it certainly seems that leaving out the likes of Jackie Stewart, Ronnie Peterson, Gilles Villeneuve, Jim Clark, Stirling Moss and Juan Manuel Fangio leaves a big hole in the final conclusions.
Or perhaps it’s just a reminder that data and numbers will only get you so far. Murray Walker – who had probably seen and commentated upon more F1 laps than any other human – always had one ready answer to the question of who was the greatest and fastest driver. Tazio Nuvolari, the great pre-second World War Italian racer, who vanquished the government-funded German teams in their own front yard at the 1935 German Grand Prix, using an apparently obsolete Alfa Romeo. That’s a victory that no bot could ever have imagined, no matter how hard it stared at the data.
- Sign up for Business push alerts and have the best news, analysis and comment delivered directly to your phone
- Find The Irish Times on WhatsApp and stay up to date
- Our Inside Business podcast is published weekly – Find the latest episode here
