Artificial intelligence should not write pregame articles for major publications

ESPN publishes daily content written by artificial intelligence.

Not only do these articles leave behind unedited flaws and factual inaccuracies, but they are found word-for-word across the internet on different sites.

Take this pre-game article from 2021 where fans on Reddit pointed out a major blunder.

The headline of the article states that Luka Doncic is supposed to lead the Dallas Mavericks into a contest with the Minnesota Timberwolves on December 19, 2021.

To fan’s dismay, Doncic never played in this game. The funny thing is, Doncic is actually listed as out with an ankle injury inside the article, an error a human would have caught.

Fans were left disappointed at the arena when they showed up to watch Doncic play, only to learn he was out for the game.

The Sports Illustrated and ABC7 websites also use these articles, which are produced from a company named Data Skrive.

Data Skrive “generates compelling, search-engine-optimized sports articles, graphics, and social content.” Yet, even with compelling statistics that are posted rapidly, the articles aren’t edited to be accurate for the game they preview.

Here’s another example, as fans sound off on this pregame article from ESPN which detailed a matchup between the Bucks and Lakers. The fans wanted to hear about Andre Drummond, yet only Lebron James and Anthony Davis were mentioned. Neither of them played in this game.

“It’s meant to help take care of mundane labor, freeing up writers to do something more valuable,” said Data Skrive CEO Brad Weitz in a geekwire interview.

Real journalism isn’t mundane labor. An article that qualifies as real journalism has research, observations, and interviews all instilled in it. These Data Skrive articles only provide one of those.

“Mundane labor” would be a poor way to describe these pre-game articles, since they provide great value to fans.

These companies need to employ editors who will fix the mistakes that the A.I. produces.

With the NBA playoffs currently going on, fan viewership is through the roof compared to season average. Fans are more invested in these games than any other time of the year.

Christmas Day, known for having the highest viewership in the NBA, averaged 4.27 million views.

A recent Lakers vs Warriors playoff game held 5.41 million viewers.

ESPN is the largest global sports network, they have more than enough money to properly pay editors for these articles.

They made approximately $4 billion in revenue last year. Their C.E.O. has a net worth of $10 million.

Sports Illustrated made about $80 million, while ABC7 made about $5 million in 2022.

The articles have good statistics that are applied quickly before every game. If ESPN were to just purchase these statistics from Data Skrive, then a human could provide deeper insight.

Current pregame articles on ESPN are very clunky. They don’t have good flow, and provide little insight beyond statistics.

Game two of the Western Conference semifinals was May 4, 2023. The automated article only compares player statistics from one team against the other, yet there are no interviews or observations.

Human input is missing!

Players have a lot to say after playoff games (mostly), which is why their insight is a valuable setup for the next game.

They talk about referees, how they’re going to defend certain offensive sets, and also how their offense can become more efficient.

“So it goes back to us having to be ultra-dialed in and locked in for game two, knowing that he’s going to come out firing. Klay [Thompson], all of those guys are going to play better. Team’s going to play better, the crowd’s going to be even louder,” said Lakers forward Anthony Davis in an interview with NBC Sports Bay Area.

As it turned out, Davis’ insight was spot on. Klay Thompson had 30 points, while shooting 8/11 from three the very next game.

These simple interviews provide great context for the previous game and also help foreshadow what’s to come. Yet, these automated articles still remain lifeless.

Worst of all, these publications don’t separate themselves from each other with their pregame content. They post the exact same automated article written by Data Skrive. Fans can’t expect a variety of opinions from these publications.

There is no point in reading these articles beyond the statistics.

Stop being cheap and lazy.