When Sports Meets AI Slop

"This sort of manufactured outrage is as profitable as it is corrosive"

I’ve been developing and teaching a class we call Criticism in Sports Media for the last two semesters. Students are learning to consume and interpret media critically, place it within broader contexts, and examine the structure and meaning of the material. This, I say, gives one an appreciation of sport media’s role in contemporary life, because sports reflect the values of a culture.

It’s a good course, and helpful. Students know there’s a lot going on, and they’re trying to understand the media landscape that surrounds and inundates us all. They are coming to understand that there are some things they don’t understand, and they’d like to try to make some sense of it.

The class spends a lot of time on the printed word and on documentaries, and we discuss social media and, lately, AI content.

Now, at the end of the term, I wanted to leave them with a lasting impression about recognizing and addressing AI. I’d been saving this piece -- co-authored by Dr. Matthew Facciani, a sociologist at Notre Dame and Dr. Beth Malow, a Vanderbilt neurologist -- which offers a highly accessible three-prong process that we can all use. I commend the read to you in full.

(U)nfortunately, this kind of content is becoming more common across nearly every major social media platform.

This is a serious problem. But it’s not a hopeless one.

To navigate this new information environment, we need to combine psychological literacy, media literacy, and policy-level change.

Facciani and Malow had videos in mind when they wrote that piece, but for class I used fake photos. The same concepts apply. This week, we’re seeing it apply in sports, specifically.

Bill McCarthy, a digital investigation journalist covering online mis- and disinformation for Agence France-Presse has been sifting through a barrage of AI slop of Cole Allen, the man now charged with attempted assassination at the White House Correspondents’ Association dinner last weekend. Various accounts have passed Allen off as a former driver for rock star Sammy Hagar, the rapper Jelly Roll, Judge Judy, journalist Aaron Parnas, Mark Wahlberg, and Bad Bunny. Others have had him as a former employee of Hendricks Motorsports in NASCAR, and in the employ of famed NASCAR driver Cole Trickle … check that … Tom Cruise.

Some wise scrollers caught on that in many of those photos -- and each post had photos, two buddies close together, more friendly than acquaintances, often with a chummy, red-carpet sort of vibe, all meant to associate this man with these other people – he was frequently wearing the same clothes. It’s a good catch. Facciani and Malow would call that vertical reading, where we are staring just a bit harder at the images, counting for extra fingers, looking for nonsensical letters, studying for flaws. Vertical reading is helpful, but it isn’t a perfect antidote.

For example, some of the slop focused on sports tie-ins. Some accounts said that Allen had worked for the Minnesota Vikings, the Chicago Bears, the Arizona Wildcats, the Houston Texans, the LSU Tigers. In his free time, he’s apparently also been working for the Celtics, Penn State, Ohio State, the Carolina Panthers, the Georgia Bulldogs, the Cleveland Browns, the Tennessee Volunteers, the Arkansas Razorbacks, Notre Dame, and and and …

In each of these he’s wearing team gear. There’s a prominent logo on the wall behind him. There’s often a lanyard or a badge. In isolation, any one of those could seem legitimate. The context clues seem right. An AI program can change the people in a photo with ease. It will manipulate clothes, as well. And as AI improves, some of the flaws will become less frequent. That’s an important reason vertical reading technique needs some help.

Facciani and Malow suggest you do what’s called lateral reading. This is looking for the same information elsewhere. Is it also on other platforms? Are people or outlets with established credibility carrying this photo or video? These are good questions to ask to protect yourself, and your feed.

You can see the collection of slop that McCarthy was collecting here.

You can read AFP’s story here.

What AI gives bad actors is a low cost, low effort product. That realization might be the point of much of our distress over this sort of slop. It’s easy to make and produce. Then there’s a key unknown that is equally bothersome. To what end? What is the point of all of this? Facciani and Malow point out the techniques involved aren’t that different than other forms of disinformation. The goals are often the same, too. The way such content tries to exploit us is similar as well: I hate Tennessee and this man obviously worked for them (look at that photo!) and that’s enough to motivate me to spread this photo of an alleged assassin in Tennessee Orange just a little bit farther.

The good news is I don’t have to know the real motivation for why someone wants me to think this man worked for all those universities and football teams and celebrities across the country. I need only know that this sort of manufactured outrage is as profitable as it is corrosive, and that this is something I might see. I need to know to be aware of it, so that I am careful to not pass along such damaging content to others.

Fortunately, my students know that, as well.