Oracle Red Bull Racing’s AI tool no longer just helps engineers look up Formula 1 rules. It now recommends what the team should do during races, executive Jack Harington told CNBC in a newly published interview. His example: advice to surrender Max Verstappen’s position to avoid a potential penalty, rather than defend it.
From finding a rule to recommending an action
Oracle and the racing team developed the strategy agent together. Its starting assignment was narrower: a search tool to help engineers interpret Formula 1 regulations and past rulings.
Harington, the team’s head of technical and performance partnerships, said the system now draws on a wider range of information sources. He described it as agentic AI—a system whose role here extends beyond finding information to recommending an action. “It’s actually making decisions for us,” he said.
The advice that went against racing instinct
To illustrate the change, Harington pointed to the 2025 Italian Grand Prix at Monza. Verstappen cut a corner while defending his position. The agent recommended giving the position back to avoid a potential penalty, he said.
That advice went against the team’s competitive instincts, Harington said, but Red Bull followed it. Verstappen subsequently overtook his rival. The race-day example predates the newly published interview: Harington was describing the agent’s use in 2025.
Formula One is and will always be a human sport
Jack Harington, Oracle Red Bull Racing, speaking to CNBC
Despite describing the agent as making decisions for the team, Harington stressed that humans remain involved. He does not expect AI to completely replace engineers and strategists. Instead, he sees it enhancing their capabilities and allowing them to focus on their strengths.
His explanation rests on the number of variables a team must weigh. Harington highlighted not only what happens on the track, but also developments off it and the broader championship picture.
Reader comments
Newest comments first. Replies stay oldest first.