Tech touchdown: How AI can predict injuries and select ticket prices

The NFL thrives on routine, so it’s no surprise that the workweek for many teams begins the same way.

On Monday morning, as coaches break down film, trainers, strength coaches and data analysts fire up the same online dashboard: Digital Athlete, an artificial intelligence platform that helps teams maximize, and even predict, player performance.

Built by data engineers, biomechanists and AI specialists, Digital Athlete includes data on more than 2000 players, each represented by a digital avatar that reflects everything they have done in practice and games.

The models can identify subtle changes in movement that may signal fatigue or higher injury risk — whether a player is ramping up too fast or pushing beyond his usual limits. Teams use that information to prevent injuries, manage recovery and ensure their rosters peak when it matters most.

In Minnesota, for example, team employees log in and study a list of players every morning throughout the season. Some names appear highlighted in red, an alert that a player’s workload or movement patterns may put him at higher risk for injury.

Trainers and sports scientists huddle, comparing the data with what they have seen on the field. Maybe a veteran skips a few drills. Maybe a wide receiver wears a no-contact jersey or a lineman sits out a Wednesday practice.

The process isn’t foolproof, but it is light-years ahead of a decade ago, when teams relied on the naked eye and an athlete’s own candor.

“What we’re trying to do is ask the question: When do we know an athlete is overworked or underprepared?” said Tyler Williams, the Vikings’ vice-president of player health and performance. “Well, we only know that by measuring the data, right?”

Across the sports world, AI has quickly moved from an emerging curiosity to an essential part of daily operations. Computer vision tools analyse pitching motions and golf swings. Wearable sensors and sleep trackers feed machine-learning models that customize recovery programs. Scouting departments use predictive analytics to find players who fit a team’s system and identify long-term injury risks. Referees and umpires rely on AI-assisted replay.

And league and team officials use AI tools to schedule games, set ticket prices and more.

Laura Harvey, coach of the National Women’s Soccer League’s Seattle Reign, said on a recent podcast she has used ChatGPT to plan strategy and tactics for some games.

“I put in, ‘What formation should you play to beat NWSL teams?’ and it spurt out every team in the league, what formation you should play,” Harvey said on the “Soccerish” podcast. “ … For two teams, it went, ‘You should play a back five,’ so I did. No joke; that’s why I did it.”

As private companies flood the AI space brimming with promises, teams, leagues and athletes are scrambling to keep pace in what has amounted to the sports world’s latest arms race.

“At the end of the day,” Williams said, “we’re all looking to be the most effective and efficient at what we’re doing to provide the resources of these players to be their best week in and week out.”

Date embrace

On the outskirts of Charlottesville, housed in a converted machinist’s shop, sits one of the most advanced sports laboratories in the US.

Inside the vast workshop, helmets are blasted by spring-loaded battering rams, cleats are tested across turf and synthetic grass, and field surfaces are studied for every kind of wear and tear.

At the centre of it all is a performance lab called Biocore, a longtime NFL health and safety partner and a subsidiary of Infinite Athlete, the data and performance company behind the NFL’s Digital Athlete program.

The NFL hasn’t always been the most progressive sports league. But it started embracing the future in 2015, when it partnered with Amazon Web Services on its Next Gen Stats.

That led to a treasure trove of data, so around 2019, Jeff Miller, the NFL’s executive vice president of player health and safety, travelled to Silicon Valley with Jeff Crandall, the head of Biocore, to explore the possibilities.

“We’d say, ‘Look, we have all this information. What would you suggest we do with it?’ ” Miller recalled.

The series of meetings led to an expanded deal with AWS and the eventual launch of Digital Athlete, now a staple for every team.

The performance lab is part gym, part television studio. A halo of high-speed cameras circles the room, four separate systems capturing every possible angle. The floor, split between field turf and polished hardwood, rests on force plates that measure every ounce of pressure an athlete applies. Each movement under the cameras can be seen in real time on nearby screens, rendered as skeletal wireframes.

“We can develop a model of an athlete that has their skeleton, has muscles, and we can actually say, ‘Based on this motion and this force, we know that his quads aren’t pushing enough— you may want to look at that,’ ” Crandall said.

Crandall is a biomechanist with a background in auto safety, and his Virginia lab is where Digital Athlete largely took shape. While the idea might have sounded like science fiction just a decade ago, the concept of a digital “twin” has seen a variety of applications outside sports, including jet engines and wind turbines. Crandall and the NFL sought to apply the same concept to football.

The models have now powered thousands of simulations for rule and field changes, including the league’s redesigned kick-off. Analysts also have used it to test playing surfaces and equipment before real athletes ever take the field.

Every tabulation happens faster than was ever thought possible. Counting helmet-to-helmet contacts once required 15-plus staffers spending days rewinding and pausing video.

Now it can be done instantaneously, with technology that analyses the video and detects where the incident occurred on the field, where the contact happened on the helmet and every other factor that preceded and followed the collision.

Much of the transformation started with Hawk-Eye, the optical-tracking system first developed for tennis and cricket and now installed in nearly every big league sports stadium or arena in the country.

Its high-speed cameras feed precise positional data — dozens of coordinates per player, per second — into Infinite Athlete’s complex engine.

“We have the capabilities to deliver the data in just a few seconds of latency — not the next day — so you can see how your athletes are performing in real time,” said Chris Brown, the chief technology officer for Infinite Athlete.

By analysing shifts in acceleration, deceleration and workload patterns captured by Hawk-Eye’s cameras, the system issues advisory alerts, not medical judgments, up to 48 hours before a game or practice. Teams can compare the AI-generated output to a player’s own performance history or league-wide data, though they only have full access to their own players.

Since the Digital Athlete team portal launched in 2023, practice-related lower-extremity strains have dropped roughly 14 percent league-wide, according to NFL data. The same technology simulated 10,000 virtual seasons to help design the league’s new kick-off, which reduced injuries even as returns increased to their highest rate in years, the league said.

Clubs are responding, logging into the system more than 5000 times last season.

“Technology is changing our lives,” NFL Commissioner Roger Goodell said during an appearance on the “Fitz and Whit” podcast last month. “It changes the way we play the game. It changes the way we watch the game. We have to embrace it. And we have to use that to make the game better.”

The cameras and AI technology are increasingly used to make split-second officiating calls, too, removing human subjectivity from razor-thin decisions such as first downs in football, goals in soccer, out of bounds in tennis or basketball or balls and strikes in baseball.

“Will AI be a part of officiating in the future? I think it will be,” Goodell said.

For Williams, in the Vikings’ team headquarters, the system works best as part of a broader conversation, not a replacement for it.

“You have digital athletes, you have global data, and you have your own team models and your clinical expertise,” he said. “Those three have to be leveraged together to make the most impactful decision for the player.”

In practice, that means managing the smallest details, whether it’s injury prevention or recovery.

“Let’s say we’re rehabbing an athlete,” Williams said. “Before, you’d have him do some sprints across the field. Well, we were just guessing, right? But now I know that player’s max speed is 20 miles per hour and he usually covers 4500 yards a session and maybe 500 ‘high-speed’ yards. So we can be more accurate and ramp him up methodically.”

AI’s ‘fine line’

The reach of this technology extends beyond football. Hawk-Eye cameras are also embedded in basketball arenas, capturing every movement on the court — 29 specific points on each player, plus the ball — and some teams contract with Infinite Athlete to make sense of the data.

In New Orleans, the technology became central to how the Pelicans managed NBA star Zion Williamson, whose explosive athleticism has long been shadowed by injuries.

Williamson, 25, began last season healthy but suffered a left hamstring strain, missing 27 games before returning in January. As he ramped up, the Pelicans closely monitored his workload using Infinite Athlete’s live-tracking tools.

Daniel Bove, the Pelicans’ director of sport science, sat near the team’s assistant coaches on the bench, armed with a tablet and monitoring Williamson’s load management “literally every minute that Zion’s on the floor,” David Griffin, the team’s former executive vice president of basketball operations, said last season.

“He can tell when Zion’s load is spiking, he can tell when he needs to be shut down, and we’ve got the ability to really monitor his progress,” he said.

Williamson, pictured, posted his first triple-double in February and his second 12 days later. He sat out six games because the team didn’t want him playing on back-to-back nights, and he averaged 25 points and more than 28 minutes per game before a bone contusion in his lower back ended his season in March.

While the NBA allows video to be reviewed on the bench, NFL rules curb what teams can access on sideline tablets — no video, no live internet connection, only still photos.

While Goodell said it can be difficult to balance “tradition versus that innovation,” many around the league believe the NFL will be among the last leagues to embrace all of AI’s in-game offerings. Across all levels, though, others are taking advantage of systems that can inform in-game strategy, play calls, line-up changes, roster construction or player management in real time.

“It seems to me we’ve only hit the tip of the iceberg in regards to how we use AI to improve performance for athletes,” Brown said.

Many in sports agree, though some tools that help keep players healthy raise questions about ownership and privacy.

How much of an athlete’s “digital self” does he control?

Who sees what and when? Agnieszka Antoszkiewicz, a former integrity manager for FIFA, draws a sharp line between officiating aids and models that assess people.

She worries about profiling and pressure — athletes labelled injury-prone or coerced into sharing intimate data. And as biometric and neurological data enter the mix, the stakes rise again.

“Using AI to detect whether a ball crossed the line is one thing,” said Antoszkiewicz, who runs sports consultant firm OrdoStrategica.

“Using it to assess an athlete’s potential or mental state is another. There’s a fine line between creating opportunity and exploiting it — especially when AI models are built on biased data.”

Still, the momentum is real. Last week, Infinite Athlete was acquired by Exos, a coaching and performance company that operates high-end training centers across the country. Exos, led by Mark Verstegen — who also oversees performance for the NFL Players Association — already works with the union and helps prepare many top draft prospects each year.

The transaction is expected to further integrate AI into football training and player development.

Brown envisions the next leap: fully connected team headquarters in which every data stream converges in real time.

Hawk-Eye’s cameras would sync seamlessly with load cells, force plates and wearable sensors, creating an uninterrupted portrait of the athlete at work — from practice field to weight room to locker room to cafeteria.

“AI scales really well with data. It needs data,” Brown said. “So the more data we collect, the more powerful AI becomes. And the more powerful AI becomes, the more powerful the data collection process becomes.”