It has been a while since I was first asked if I was afraid that artificial intelligence would take over my job as a writer. And since that question, we’ve all seen new developments in AI on the daily. But most don’t warrant, in my opinion, my opinion.

Until this. Last month, AI debuted at the Aspen, Colorado Winter X Games. Now, it would be quite normal to hear that AI assisted in making the schedule or helped keep the broadcast running smoothly. But that wasn’t the news.

It judged the damn competition.

With a series of cameras, an AI watched each run and spat out a score at the end. Now it’s important to note that the AI given scores were not used in the judges decision making. It was only a trial run. A trial run that leads to future adoption? That’s unsure, but it’s clearly the direction we’re moving in.

And moving quickly. According to a New York Times article, different AI technologies have already taken a role in tennis line judging, soccer offside rulings, and baseball strike zone monitoring. We’re skaters though, so fuck all that. Let’s talk about how this could impact the future of judging within the love it or hate it world of competition skateboarding.

Eric Koston, X Games 2000

Objectivity. It means freedom from bias and absence of favoritism. It is one of the highly touted advantages when the conversation of AI judging comes up. But, is total objectivity what we want in a skate judge?

Most sports work in absolutes. A strike is a strike because it is within the strike zone. It’s black and white. A kickflip isn’t absolute. It could be mobbed and sketchy in a bad way. Shit, it could be mobbed and sketchy in a good way. It could also be caught perfectly and have that slight downward tweak that makes my heart happy.

This is to say that something as simple as a kickflip can vary widely. It has no strike zone. It is subjective, and that makes it especially difficult to judge. This subjectivity is one of the main pillars of anti-competition arguments. And understandably so. How do you judge something based on opinion?

If one were to turn the competition into an objective format for AI they’d have to dissect and distill all skate tricks into statistics.

Pat Stiener, Tampa Pro Judge, was quick to find issues in breaking skateboarding into statistics. “Is a feeble grind better than a smith grind, or are they equal? Then you would have to add in whether it was switch or nollie, even though some tricks in my opinion are better than their switch counterparts like a frontside bigspin is harder/cooler than a switch frontside bigspin. Same can be said with fakie and switch, a fakie crook down the rail is harder than a nollie crook and less played out.”

Expanding on a similar sentiment is Scott “Big Cat” Pfaff, longtime judge of Street League Skateboarding events, “This could lead to a bunch of skaters gaming the system by all learning the same five tricks because that’s what some algorithm has determined are the “best.” It would turn skate contests into figure skating.”

And looking at how the recent snowboarding competition was judged by AI, you’ll start to get a feel for what that means. CEO of X Games Jeremy Bloom explained to NPR that the machine used “amplitude [height], difficulty of the trick and the precision of the landing,” as criteria for its judging.

Applying this criteria to skateboarding, a score would be given based on size/length of obstacle skated, objective difficulty of the trick, and the cleanness of the rollaway. This essentially turns the contest into a bastardization of track and field, of longest, fastest, highest.

While those are the objective measurables in skateboarding, we all know there’s more to it than that.

Dylan Rieder, Street League 2013

Alright y’all, here we go. The meat and mashed potatoes of the conversation. Can AI understand style and its importance to skateboarding?

Jason Rothmeyer, Head Judge for The Boardr, makes it pretty simple, “ Clearly a run on paper vs a run actually done are two different things. If you go out there all knocked knee’d doing ugly ass tricks like double flip to boardslide and crooks to back lip, is an AI going to score you high because you “killed it” with combos but in reality we’re all throwing up in our mouth watching you skate?”

Looking good on a skateboard is impossible to quantify or look at objectively. Sure, an AI may be capable of sniffing out a sketchy landing in regards to style, but I highly doubt it would agree, given roll away and trick difficulty, that Dylan Reider’s gap back smith in a 2013 Street League event was one of the best tricks done in the contest that year (and that’s just my opinion).

As Big Cat would put it, “Why can you watch 100 people do the exact same trick and have one of them stand out?”

The delicate balance between the outright difficulty of a trick and the way in which you did it has always been a question when considering not just competition skateboarding, but skateboarding at large.

And while I admit it could be funny to see Tony Hawk Pro Skater ledge combos at a competition level, that’s probably not what we want. Insert age old adage, “Just because you can, doesn’t mean you should.”

Subjectivity in judging, not for the purpose of rooting for a certain skater but for upholding the stylistic integrity of good skateboarding, is important. It’s these opinions, this context, that human judges bring.

Stiener had this to say about style, “If [Dennis] Busenitz is hauling ass doing a back tail on a ledge and pops out, that would be better than another skater going super slow doing a front crook. AI would probably have a front crook scored higher than a back tail though.”

At the end of the day, a computer lacks taste.

The MIT Technology Review dove into this topic in regards to gymnastics, which happens to be another subjective sport, and had this to say, “Nadia Comaneci’s first “perfect” 10 at the 1976 Olympics wasn’t perfect; she shuffled her feet on her landing. But the routine went down in gymnastics history thanks to creative judging, a reward for a certain je ne sais quoi Comaneci brought to the mat.”

It’s that je ne sais quoi that objective judging probably wouldn’t consider. And for skateboarding, that means ignoring the style, the ever important trick selection and creativity that draws most of us to the activity in the first place.

But, one of the most compelling, albeit scary, aspects is that AIs learning capabilities are endless. It’s not grasping at straws to say that one day AI could ingest every opinion ever aired online about skateboarding, from deep SLAP sentiments to what Burberry Erry thinks.

In doing so, AI would have access to the collective consciousness of skateboarders. And if that happens, I suppose AI could begin to try and decipher style. That’s quite the homework assignment though, and definitely not what we are considering here today.

Dennis Busenitz, Tampa Pro 2011

A computer doesn’t tire. It doesn’t blink. It doesn’t need to take a coffee shit at inopportune times. It’s these problems, amongst others, that AI in judging might be able to help assist with.

“This technology can see way more than the human eye. And the technology can index every run that’s ever been done before,” said Bloom.

Big Cat said something similar, about AI being able to potentially “catch small details that a person might miss, clock speed, distance, the way someone does or doesn’t lock into a trick, etc.” The things you would need a magnifying glass or measuring tape to figure out, you could have right by your side.

Another benefit of AI could be the speed in which it produces a score. Judges are asked to have a score within seconds of a run ending, which can be a daunting task.

“A lot of times at live events we are pressed to give a score very quickly, I’m talking 2-3 seconds after a run is over with almost 0 time to think. I always tell judges that you basically have to have a score in your head BEFORE the run is over and then when that last trick is landed/bailed, adjust accordingly,” said Rothmeyer.

That’s a lot to ask from a person, especially considering how they are supposed to judge each run against every other done that day. At Tampa Pro there are 30+ skaters doing two runs each, putting the total number of runs to consider above sixty. And that’s in the semi-finals alone.

Maybe AI could help with some of the data crunching.

And amongst all this data, human judges have to try to not get swayed by histories, narratives and crowd opinions. I’m not claiming big Scientology is out there paying off the judges for Cariuma skaters, but I’d be awestruck if narratives and ratings don’t, at the very least, play a subconscious role in decision making.

Theoretically, AI would consider none of this, and if used correctly, that level of objectivity may be a tool for the judges to use.

If we are to continue down this line of thinking on machine objectivity, AI could remove bias against certain skateboarders as well. Think about the Tonya Harding predicament. I don’t think any skater is currently getting bent over the barrel, but we could assume if AI was judging that it doesn’t have a favorite skater.

Yuto Horigome, Paris Olympics 2024

To take the easy road would be to shake this whole conversation off as future dystopian bullshit, but that’s doing it a disservice. Since skateboarding was adopted into the Olympics, one could say competition skateboarding is at a new cusp.

And we’ve seen it develop in real time. World Skate competition circuits, qualifying point systems, and new companies involved in skateboarding are all emblematic of skateboarding bending to the desires of the Olympics.

This is all to say that if Olympic rulings, which have already decided to embrace AI in some aspects via their Olympic AI Agenda, decide that it would benefit skateboarding judging, then it has the potential to happen.

So we need to look for a proper balance. Sooner than later.

This balance will inevitably land in the grey area between objectivity and subjectivity, but I could imagine a world where AI helps judge objective measures like speed, number of tricks landed, and obstacles used, while human judges focus on style and creativity.

“Maybe it would be nice to see an AI suggested score on your iPad or something just to fact check yourself to make sure you didn’t miss anything,” admitted Stiener.

But the questions remain: How do we denote percentages to aspects of skateboarding like style and difficulty? And how important is style?

For some, it could be everything. For others, they want to see skateboarding pushed to its limits, style be damned. Regardless of opinion, Rothmeyer understands, “AI is out there, and everyone wants to see how exactly it will perform for them and what things it can help with. It’s a natural progression.”

But he goes on, “At the end of the day, I don’t think it’ll be nearly as effective as a properly seasoned judge.”

Written by: Ben Komins
Header illustration: Wilson Lucas
Special thanks: Pat Stiener, Jason Rothmeyer, and Scott “Big Cat” Pfaff
Ask Meta AI on Facebook
Confuse an 80-year-old on Instagram and Twitter