Humans teaching AI: How driverless cars racing in Abu Dhabi run on code, brainpower

From ex-soldiers-turned-AI engineers to mechanics re-learning motorsport, the people behind A2RL prove “autonomous” doesn’t mean human-free. In the garages at Yas Marina Circuit, the air smells of tyre rubber and solder. Here, in the so-called driverless racing league, the real competition unfolds long before the lights go green.

“The physical car itself is the same, but all the autonomous control systems have all been updated and all the wiring has evolved,” said chief mechanic Michael Noss, who oversees four of the A2RL cars. “Last year a lot of the cars could drive around slowly at 20 to 30 kmph; now they’re not far off what a human would do.”

Recommended For You

CentFX announces participation in iFX Expo 2026 as silver sponsor

Oil outlook 2026: Surplus keeps Brent near $58, but long-term supply risks loom

AgraME rebrands as GROWTECH. MIDDLE EAST, uniting regional legacy with global agri-tech power

Time magazine names 'Architects of AI' as Person of the Year

Adnoc Distribution adopts AE Coin payments across its retail network

 

Unlike Formula One, there are no pit-to-driver radios or tyre gambles. Each car races alone — making its own decisions based on sensors, algorithms and the personality its engineers have coded into it. Yet behind the software stands a team of human specialists who keep those machines alive.

Stay up to date with the latest news. Follow KT on WhatsApp Channels.

Mechanic Arshed T.S. described the learning curve of working on a vehicle that thinks for itself. "I had to learn how to change the electrical systems — radar, LiDAR and cameras."

“Autonomous can be more fun as well,” added his colleague Amiru Syakir.

The same blend of curiosity and technical obsession drives the teams competing this weekend, as the world’s largest autonomous car race takes over Yas Marina Circuit on November 15, with 11 teams from around the world fighting for a $2.25 million prize pool. According to organisers, six fully autonomous cars — from teams representing Germany, Italy and the UAE — have qualified for the Grand Final, while five more, including Code 19, will compete in the Silver Race.

From Italy’s TII Racing, technical director Giovanni Pau said his engineers are “testing our algorithms to try to optimise according to the track… you want to be fast and to get fast very quickly.” The team, which joined the league this year, has already reached top speeds of 248 km/h on the back straight.

Sometimes, though, even a perfect line of code meets a human-style surprise. “Sometimes, you expect a performance, and you get something a little better. In a bad way, sometimes you think everything is going to go perfect, and the car stops in the middle of the track,” Pau said. “You have those three seconds — ‘don’t touch the wall, don’t touch the wall.’”

For Abu Dhabi audiences who saw last year’s human-versus-AI demonstration, Pau’s team was among the pioneers. “We did the demo with Daniil Kvyat last year,” he recalled. “He was about ten seconds faster… but you just need time to learn.”

That challenge will return this weekend, when reigning champions TUM face former F1 driver Daniil Kvyat in a fresh Human vs AI showdown, a headline act for A2RL Season 2’s Grand Final. Organisers say the event marks “the first time that six fully autonomous race cars will compete wheel-to-wheel on track.”

Across the pitlane, the US-based Code 19 team brings a very different story. Co-founder Lawrence Walter and his partner Oliver Wells came to motorsport from the military. “We saw how important robotic and autonomous systems were becoming in the military, and we knew that pushing this technology on a racetrack was going to have huge potential,” Walter explained. “Both my co-founder and I were huge racing fans.”

A former US Army intelligence officer, Walter served in cyber and special operations. His partner, Wells, remains in the Army Reserve. Between deployments, they built what they call the world’s first professional autonomous racing team.

“The league took a risk on us as having no experience in motorsports,” Walter said. “But we’ve proven our value to the league by what we’ve been able to do on track.” For them, race preparation feels more like military fieldwork than a weekend sport. “It’s all about preparation and speed — not just speed on the track, but speed of engineering and speed of innovation,” Walter said. “You go out for 15 minutes, find problems, come back, take gigabytes of data from the car, fix it, and go out again.”

Failure, he added, is part of the process. “We’ve had several crashes — but that’s how you learn. You can’t be afraid of failure. The key is how you learn from those failures.” This season, Code 19 has focused on localisation — how precisely the car knows where it is on the circuit. “GPS will get you within a metre, but we’ve developed algorithms to get sub-centimetre accuracy,” Walter said. “If you don’t know where you are, you can’t reliably go as close to the corners as possible.”

Walter’s team was last year’s top rookie, finishing fifth out of eight. Now, with 11 cars on the grid, they aim higher. “The fastest AI lap is 58 seconds — the fastest human so far is 59.” This autonomous driving record was faster than the benchmark set by a professional racing driver during data collection laps. That near-parity is what excites the pit crews most. As the machines grow quicker, the human role shifts from steering wheels to strategy boards, from driver instinct to digital precision.

“You have to trust what you’ve coded, what you’ve built, but also what happens live in that one lap,” said Pau. When the autonomous race begins at Yas Marina on Saturday, it will be silent inside the cockpits. Yet in the garages, every human heartbeat will be racing — listening to the hum of servers and watching the world’s fastest experiment in machine intelligence unfold on track.

• Abu Dhabi

Haneen Dajani

haneen@khaleejtimes.com

Haneen Dajani is special correspondent in Abu Dhabi with over 15 years of reporting experience. She’s also a passionate athlete, full Ironman finisher, and mountaineer who loves to embark on unusual challenges.