Wimbledon’s AI Announcer Was Inevitable

The Wimbledon announcer sounds a little like Helen Mirren if she’d just been hit with a polo mallet. I’m watching match highlights between Ons Jabeur and Magdalena Fręch on the tournament’s website when a voice says, “Jabeur, from Tunisia, will play Fręch, from Poland, on the renowned No. 1 court in the first round.” Fręch is mispronounced, as is Tunisia, and the word renowned is used oddly dispassionately, as if it were being repeated for a competitor at a spelling bee.

This is a commentary chatbot, introduced with considerable fanfare at the All England Club this year. Another version, a “male” voice, sounds like your uncle from Queens trying to do a Hugh Grant impression. These AI commentators provide “play-by-play narration” for highlight reels published online. They are the result of a partnership between the All England Club and its longtime corporate sponsor IBM, which has been part of Wimbledon for so long that it introduced the “Data Entry Keypad” back when John McEnroe and Martina Navratilova were still playing.

Nothing about the technology feels ready for prime time, and it seems that Wimbledon and IBM know it. It’s limited to short clips, and the feature is nearly hidden on Wimbledon’s site—it requires you to click on a video, then click on tiny headphones in the corner, then choose “AI Commentary” mode. But IBM explicitly said in its Wimbledon press release that the goal is to eventually have AI commentary during some of the actual matches, and the European Broadcasting Union has “cloned” the voice of a major commentator that it plans to deploy via AI at an upcoming event. This is, in many ways, the logical direction that the world of sports broadcasting has been trending toward for many years now. Your favorite sporting events are not yet broadcast by robots, but it sometimes seems like they already are.

Powered by a large language model that is trained on specific Wimbledon terminology such as gentlemen’s draw instead of men’s draw, the bot first identifies highlight-worthy videos based on factors including crowd noise and player fist pumps. Gregor Hastings, an IBM spokesperson, said in an email that “the commentary is designed to be simple and not too overbearing,” but the end result—a few bland phrases at the conclusion of some points—is painfully boring.

That is especially true when compared with the prickly observations of, say, McEnroe, the famously tempestuous former player who has long been a broadcaster for ESPN. McEnroe’s act has cooled in recent years (when was the last time you heard a classic “You cannot be serious”?), but he’s still far more animated than his younger colleague at Wimbledon, Chris Fowler—a perfectly professional announcer whose calm cadence is interchangeable between college football and tennis and even hockey. This mirrors what has happened in sports broadcasting in general.

In the early, formative days of sports television—think the 1960s and ’70s—the broadcaster himself (and it was almost always himself) was the star. And why wouldn’t he have been? He was, after all, the only one talking. For years, broadcasters were as much showmen as they were commentators. They had distinctive styles and clearly defined personalities, whether Mel Allen or Bob Costas in baseball or John Madden or Pat Summerall in football. Much of this evolved from sports television’s biggest star of all, Howard Cosell, whose elliptical style and brashly outspoken nature made him so famous that he regularly outshined the games and players themselves. He starred in a Woody Allen movie and became a bit of a national conscience, famously telling the world on Monday Night Football that John Lennon had been killed.

It certainly helped that for so long, the audio and visuals of a sports broadcast were so lacking that you needed a voice to guide you through games. In today’s HD age, it’s remarkable to watch old highlights and realize just how difficult it is to decipher what’s going on. This was the prevailing mindset of sports television for decades, to the point that Monday Night Football hired the comedian Dennis Miller as a color analyst even though, until his first assignment, Miller had never attended a football game in person.

This also led to a spate of local-sports-broadcast personalities who would become more synonymous with their teams than the players on the field were, including the Cubs’ Harry Caray, the Phillies’ Harry Kalas, the Dodgers’ Vin Scully, and the Cardinals’ Joe Buck (as a Cardinals fan, I think I’ve heard Buck’s voice more in my life than I have my father’s). These regional broadcasters, because of their ties to the community and their ubiquity, became local treasures, even civic institutions: When you turned on the radio and heard Ernie Harwell’s voice, you knew you were in Detroit. Caray became such a Chicago staple that the Chicago Sports Museum is adjacent to Harry Caray’s restaurant, and Caray’s statue outside Wrigley Field is perhaps more famous than any player’s statue that is out there. Even if you didn’t watch sports regularly, if you lived in these towns, these were your local celebrities.

But in the past decade or so, this has begun to change. These beloved older announcers, as they have retired or died, have been replaced by bland, safe, mostly personality-free talking heads hired specifically to be noticed as little as possible. Each one sounds the same, down to the same accent. This makes sense when you consider the transitory nature of broadcasting itself: You don’t want to embrace that yinzer accent when you might get transferred to Topeka next week. In an age when networks are terrified of a broadcaster saying something that will get them pilloried on social media, the job has become anodyne and hopelessly watered down.

But the transition to AI announcers is even more of a function of most networks’ corporate contracts. National broadcast entities such as ESPN and Fox have billion-dollar deals with the sports leagues, which have control over who is selling their games to the wider public. More and more, regional networks such as YES Network and Bally Sports (which is now bankrupt) are partly owned by the teams themselves, which have zero interest in some wild card going rogue and criticizing their team—a battle that has been going on for a long time. If you are a broadcaster for, say, the Yankees, the word Yankees is right there on your paycheck in a way it likely wasn’t in the past.

You are incentivized not to stand out—and not to hurt the brand. Why not just take the next step and employ bots? This is particularly pertinent for a tournament such as Wimbledon, a famously stuffy event that is notoriously controlling of its image, to the point that until this year, women players had to wear white underwear (the men still do). If ever there was a tournament that would want to program precisely what its broadcasters say—to have an AI bot do exactly what it is told—it would be Wimbledon.

Sure, it’s difficult to imagine the Super Bowl being broadcast by chatbots. But for smaller events or more niche sports, AI sure looks like a quick and easy way to cut costs—to make telecasts, in that corporate parlance, “more efficient.” If networks think you won’t be able to tell or just won’t care, why would they pay an actual human? Witness, most recently, ESPN’s sweeping layoffs just last week, which cut some of the biggest names in sports broadcasting, including Jeff Van Gundy, Suzy Kolber, Jalen Rose, and many others. The networks have become less partners of the leagues and more subjects entirely at the mercy of leagues’ whims and their stomach for risk management. Robots, in the end, are much more easily controlled.

But perhaps all of these networks and tournaments are missing an overarching and vital point: Fans want to be upset by announcers. Cosell became such a huge star because fans were consistently infuriated by him. (The NFL commissioner Pete Rozelle once compared him to Attila the Hun.) Having a personality in the booth—or, jeez, you know, a person—provides a connection to the sport that every fan craves, even if it’s a negative connection. Even if it makes us mad. Especially if it makes us mad.

Just consider Peacock’s experiment last year. For one of the MLB games it shows on Sunday mornings, the network showed a game that had no broadcasters at all. It is easy to see, particularly if you’ve been annoyed by broadcasters before, why you might find this calming, even soothing in a zen way—just ballpark sounds, no pomp and circumstance. But fans hated it. They found it unnerving, off-putting, and confusing: Without a person watching the game with them, it was difficult to tell what was going on. The game stops becoming a sport; it’s just an activity. It’s just something else happening, in the face of an indifferent Nature.

This is the danger of AI that I suspect Wimbledon is missing: Without people to tell us how it feels to be in the presence of the greatest tennis players in the world, viewers have a difficult time finding it particularly important to be invested at all. AI broadcasters aren’t there, aren’t watching, and don’t care. So why, in the end, should we?