AI camera operator repeatedly mistakes bald head for soccer ball during live stream

AI researchers sometimes call machine learning technology “fragile.” By this, they mean that artificial intelligence lacks human understanding of the complexities of the real world, with the end result being that AI systems sometimes break in quick and unexpected ways.

For a more practical illustration of what ‘frailty’ means, just consider the case of the AI ​​camera operator deployed by Scottish football team Inverness Caledonian Thistle FC. When the pandemic prevented fans from attending matches, the club announced that they would broadcast their matches live, using a automatic camera system with “built-in AI ball tracking technology” to make sure people always get the best view of the action.

But, during a recent live stream of matches, the AI ​​camera operator repeatedly mistook (and to great effect) a linesman’s bald head for the soccer ball itself. In a compilation video below featuring a excellent soundtrack, you can see the end result. It’s as if the AI ​​has a crush on the linesman and can’t help but zoom in on his shiny handsome bonce.

Pixellot, the company that makes the camera technology used by Inverness Caledonian Thistle, has confirmed The edge that the problem was caused by visual similarities between the linesman’s head and the soccer ball. They noted that the angle of the camera did not help, as it made the linesman’s head appear to be inside the boundaries of the field and the ball itself was yellow, which added to the confusion. The company said the error was fixed shortly after completing the game.

“Last month we produced 105,000 hours of live sports without any staff on site. That’s about 50,000 games,” the spokesperson said. “In one of the games, there was indeed a tracking issue which has been resolved.”

But again, that’s why AI is sometimes called brittle. A human would never make this mistake, but a machine learning system with a limited understanding of soccer balls and human anatomy would. And while this particular example is a lot of fun, these kinds of errors can happen to any AI system. This mistake is eerily similar to, for example, some mistakes made by self-driving cars that have caused fatal crashes. In one case, Tesla’s “autopilot” software failed to distinguish between the white side of a truck and the sky.

According to tweets from the match itself (seen Going through IFLScience), the situation was so bad that the game’s commentator had to apologize for the camera malfunction at one point. One fan even said he missed seeing his team score because of the AI’s lack of focus.

Updated Nov. 5 at 4:30 a.m. ET: History updated with confirmation from Pixellot on the nature of the error.