As it celebrates more than 30 years partnering with the USTA, learn more about the AI-powered commentary and analysis features IBM rolled out for the final Grand Slam of the year.
To enhance the digital experience at the 2023 US Open Tennis Championships in New York, IBM announced new artificial intelligence-powered features for the more than 10 million fans who use USOpen.org and the US Open app.
The tech giant’s work at the annual tennis event at Flushing Meadows has come a long way since USOpen.org first launched in 1995, followed by the tournament’s mobile app launch in 2009, and IBM has been there the whole way through as the official technology partner of the United States Tennis Association (USTA) for more than 30 years. For 2023, IBM debuted generative AI spoken commentary and AI Draw Analysis to US Open digital platforms.
The company first introduced similar AI-powered offerings at the 2023 Masters in April and has since updated these features to offer the same across tennis to bring fans closer to the game.
“Many people will say, ‘Why are you guys sponsoring a tennis tournament or doing technology at a tennis tournament?’ The reality is it’s a business just like every other business on the planet,” Noah Syken, IBM’s VP of Sports and Entertainment Partnerships, said during an exclusive media roundtable. “We think of these relationships as delivering mission-critical capabilities.”
I was fortunate enough to tour the IBM Data Center at the USTA Billie Jean King National Tennis Center during the US Open this year. Joining the tour were Syken; Monica Ellingson, who leads the Sports and Entertainment Practice within the IBM iX studio; Brian Ryerson, USTA senior director of digital strategy; Maria Sharapova, 2006 US Open champion; and more IBM experts.
Here’s what Boardroom learned about IBM’s new digital offerings at the US Open.
Sign up for Tech Talk, where Boardroom’s Michelai Graham breaks down the latest news from the world’s biggest tech companies and the future of industry-shaping trends like AI.
IBM at the US Open
IBM has long-standing partnerships with The Masters, Wimbledon, and the US Open to bring emerging tech to the forefront of these major events and enhance the fan experience. This includes providing player data, rankings, predictions, highlights, and, most recently, AI-powered commentary. IBM’s data operations behind these annual events are massive and require significant computing power in order to deliver experiences almost in real time.
Each year, IBM works closely with its partners to add new layers to these evolving digital experiences, capturing and analyzing millions of data points to power them — IBM’s data center at the US Open collects 56 data points alone for every single point of a tennis match. The company operates across a hybrid cloud infrastructure in a bunker-like room on the bottom level of Arthur Ashe Stadium staffed by a team of experts and endless computer screens streaming live match feeds.
AI on the Court
One of IBM’s newest offerings at the US Open includes an AI commentary feature powered by IBM watsonx, their flagship AI and data platform that automates the production of detailed audio narration and caption that pair with highlight videos.
IBM worked closely with the USTA to build and train AI models primed to output true tennis language. During the media tour, Ellingson said AI commentary is produced within 10 to 15 minutes for all match highlights distributed across the US Open’s digital platforms. Each highlight is queued up with AI commentary and captions, is matched with accompanying graphics, and goes through a final spot check before hitting the platform. IBM leaders said it used to take hours to pull this content together, and only for select highlights.
Now, all of that has changed.
Unlike some sports commentators you’re liable to witness on television, IBM’s AI commentary is designed to be fully unbiased, sticking purely to the facts of what’s happening on the court. Still, Ellingson noted that IBM isn’t working to phase out the iconic voices fans are familiar with, but is rather focused on making commentary more accessible across all matches. Most importantly, IBM is solving a “problem of scale,” Ellingson said, in providing commentary on highlights that otherwise wouldn’t have them.
“The AI Commentary feature was developed based on the watsonx enterprise-ready data and AI platform, designed to enable businesses to accelerate generative AI work by leveraging foundation models and machine learning in one place, with their own data,” Jonathan Adashek, SVP of Communications and Marketing at IBM, said in a press release.
IBM also brought back its Power Index this year, a platform that uses advanced data analytics and AI to analyze recent player performance metrics and expert opinion to create a mass rankings index. IBM’s Power Index varies from ATP and WTA rankings because it follows player momentum from every tournament to every match, while those traditional global rankings only account for 52 weeks of data at a time.
IBM launched its Power Index to provide a new perspective on the sport while leveraging AI to pull in data and information from various sources, including the news. For instance, IBM’s Power Index puts Coco Gauff at No. 1, while her WTA Ranking is #6. Similarly, Ons Jabeur is ranked No. 2 on the index, while her WTA rank is No. 5. On the men’s side, Novak Djokovic is ranked No. 1 in the Power Index and No. 2 in the ATP Ranking. Carlos Alcaraz and Jannik Sinner take over the top three at No. 2 and No. 3 on the index, respectively, while Alcaraz is No. 1 on the ATP Raking and Sinner is sixth.
Also powered by IBM watsonx, IBM’s AI Draw Analysis feature leverages data from the IBM Power Index and Likelihood to Win features to determine the level of advantage or disadvantage for all players in the singles draw across all seventeen courts. This feature leverages both structured and unstructured data that will change as the tournament progresses and players’ matchups are set. The IBM AI Draw Analysis is updated daily through the tournament, which runs through Sept. 10.
“Over the process of the US Open Tournament, we will process 2.7 million individual pieces of match data. Then, there’s unstructured data,” Ellingson said during the press tour. “We process millions of articles, blogs, [and] subtitles from a variety of different sports-related content sources, and we use AI to extract those insights.”
One of the most notable new AI-powered features is IBM’s projected paths to the finals for each player. Users can view players’ respective likelihoods to advance past each round and their favorability to win against potential opponents.
Here’s a look at projected paths to finals for three top players on the tournament’s men’s and women’s sides, provided by IBM’s PR team:
- Women’s draw: Jabeur has a favorable draw leading up to the quarterfinals, which lines up with the IBM Draw Analysis, where she’s ranked No. 2. However, the competition gets tougher in the quarterfinals. Similarly, Jessica Pegula is predicted to have a relatively easy path to the semifinals, with expected victories in all rounds leading up to that point, including a win over Marketa Vondrousova. However, she may face Jabeur in the semifinals and potentially Gauff in the finals.
- Men’s draw: Djokovic has the most favorable draw to reach the finals, while Alcaraz faces a more challenging path to the final. He might have to compete against players like Sinner and Alexander Zverev, and then potentially Daniil Medvedev to reach the final.
The Power Index may have it right, but IBM knows there are always surprises and upsets at the US Open, so the tech giant leveraged its index data to create a pre-tournament list of players to keep an eye on. IBM’s Ones to Watch list calls out 10 players that are five or more positions higher on the Power Index than their current tour rank — from where the tournament stands at the time of writing, all of these players are still in the running. Here’s IBM’s Ones to Watch:
Women’s draw
- Elina Svitolina
- Linda Noskova
- Elisabetta Cocciaretto
- Liudmila Samsonova
- Quinwen Zheng
Men’s draw
- Max Purcell
- Andy Murray
- Grigor Dimitrov
- Alex de Minaur
- Tommy Paul
The Future of AI Tech in Sports
Tennis icon Maria Sharapova joined the press on the tour to talk about how AI tech, and emerging technology in general, is transforming access to sports data. She spoke about how different her game could have been if she and her team were able to comb through player data ahead of matches. IBM execs explained that today, athletes and coaches can not just access richer, more detailed data if they want to dive deeper into the numbers on their opponents, but can do so essentially in real time.
Sharapova, who competed professionally from 2001 to 2020, also touched on the overall digital transformation of the sport, noting an evolution far beyond the days in which fans couldn’t even watch practices. Now, these sessions are streamed online.
“I remember speaking to my manager on whether I needed a social media account at the time,” the career Grand Slam winner said. “There is a part of sport, and tennis in particular, where it’s down to heritage and tradition. Now, I’m on the other side from being a player to being a fan and loving the game so much and wanting more access than just the player’s plating the hour-and-a-half or two-hour match. By having these tools in front of me, it’s such a quick way of accessing information and just makes it a better experience.”
One thing IBM is working on mitigating in the months and years to come? How AI models “hallucinate” and drift from the tasks they should focus on. The company can do this by making sure input prompts are as clear and consistent as possible across the board and making sure data sets are updated often in order to correct errors. For instance, if there is a gap in information within a data set, the AI model only knows to pull from what it has, so it may come up with some output that doesn’t align with what a user is looking for. For this reason, IBM will continue to have human experts fact-checking and keeping up quality control as much as possible.
The tech company created its AI commentary platform specifically to celebrate the game and players, so you won’t find negative commentary here, but that doesn’t mean the goal is to sound emotionless or mechanical. In terms of evolving this feature, Ellingson said there are opportunities to bring celebrity personalities and tennis voices to the platform, and these efforts are currently in the works. Imagine watching tennis highlights voiced by your favorite tennis players — wouldn’t that be the dream? Sharapova joked that she’d be down to help train IBM’s platform on her voice for commentary purposes, all from her couch.
Tyler Sidell, Technical Program Director at IBM, said during the tour that the company is also working with the USTA to expand AI commentary and captions to more languages. Ellingson said it would be exciting if IBM could expand it into some sort of radio-style program to help fans follow along with what’s happening across all US Open courts simultaneously without even having to watch individual highlights.
Today, IBM’s AI-powered features are only available within the platforms they build in partnership with the sporting event organizers they work with. This means we may not see IBM’s cutting-edge offerings at the NFL or NBA Playoffs this season, but leaders at the company said it may be possible one day as they work to expand these sought-after features.
There is clearly a lot of wiggle room to fit AI tech into the growing sports landscape, especially at scale. It’s inevitable that similar tech and fan engagement offerings will be introduced across all sports, but for now, the US Open app and USOpen.org provide an intriguing glimpse into the future of fan engagement.
Read More:
