Aug 14, 2019 - Technology

The creation of sports highlights is being automated

Illustration of a robot with its eyes glowing and projecting light.

Illustration: Aïda Amer/Axios

When I worked at ESPN, one of my jobs was to watch a live sporting event, log everything that happened in real-time, and produce a highlight (usually one short version, one long) that would run on "SportsCenter."

The state of play: Deciding what plays and replay angles to include in my 60-second retelling of a 48-minute basketball game felt like a very "human" task. But fast forward just a few years, and that job, like so many others, is now being automated.

Driving the news: WSC Sports, a Tel Aviv-based startup developing AI-powered sports video creation tools, recently closed a $23 million Series C round to bring the company's total funding to $39 million.

How it works: WSC Sports' AI technology analyzes sports broadcasts, identifies key plays, and combines them into highlight reels in near real-time. Those short-form clips are then published and distributed at scale for clients like the NBA and Bleacher Report.

  • The AI listens for crowd noise and understands that the louder the crowd, the more important the play was. It knows that multiple replays mean something big just happened, and it’s always aware of the score, the time and the situation.
  • As a result, it can construct a narrative of the game as the action unfolds and instantly produce a highlight reel recapping what went down.

Why it matters: This technology could revolutionize how sports networks and media outlets produce highlights, which remain a huge part of every sport fan's diet even if the glory days of Stuart Scott and "boo-yah" are over.

  • In fact, while some may argue that the sports highlight show is dead, highlights themselves are more ubiquitous than ever thanks to social media.

The big picture: In addition to taking over highlight-making duties, our robot overlords are also writing game previews and recaps — and that's actually been going on for a while now.

  • Newsflash: If you read a preview of a college basketball game or a recap of a minor league baseball game this year, there’s a good chance it was written by AI designed to turn data into natural language, rather than a human being assigned to cover that game.

Go deeper: The coming impact of automation

Go deeper