What if I told you that most of the sports stories you’ll read in the future will be written by robots?
Well, actually, you probably already have perused a few AI-generated articles on athletics and didn’t even realize it.
Who Wrote It… A Person or a Program?
Read this item about the University of Hawaii baseball team:
The Hawaii baseball team pounded out 15 hits en route to a 13-6 rout of UC Davis today at Les Murakami Stadium.
By winning two of this three-game set, the Rainbow Warriors won a series for the second time this season.
Daylen Calicdan went 5-for-5 and catcher Tyler Murray was 3-for-3 to help the ’Bows improve to 13-17 overall and 3-3 in the Big West. The Aggies fell to 10-14 and 3-3.
The ’Bows seized a 3-0 lead in the second inning. After the Aggies scored single runs in the fifth and sixth innings, the ’Bows answered with three in the bottom of the sixth and five in the seventh. In that three-run sixth, Calicdan, Dallas Duarte, Murray and Tyler Best hit back-to-back-to-back-to-back doubles.
In 6 2/3 innings, Logan Pouelsen spaced eight hits but did not issue a walk to improve to 2-0.
Now, read this one from the University of California Davis:
Senior Garret Kelly tied things up in the eighth with his sixth homer of the season, while redshirt junior Logan Denholm plated the go-ahead run in the ninth with his second hit of the afternoon, helping UC Davis clinch its first-ever series win over Cal State Fullerton in program history thanks to a thrilling 3-2 comeback victory on Sunday afternoon at Goodwin Field.
Sunday’s win marked the Aggies’ sixth all-time win over the Titans to date as well as the fifth road win at Goodwin Field.
The Big West triumph pushes UC Davis’ overall record to 17-24, including 8-10 in conference play. Cal State Fullerton, on the other hand, falls to 19-24, 6-9 in the Big West. Junior Steve Ouellette grabbed his first career win for the Aggies, bringing his record to 1-3. Freshman Kyle Luckham took the loss, falling to 3-5 on the season.
Which is the human-written one, and which is the product of artificial intelligence algorithms?
Can’t tell? That’s the point. The answer is at the end of the article.
AI: Artificial Intelligence or Athletic Insight?
The sports world will never be the same now that the Associated Press has tapped Data Skrive, a Seattle-based artificial intelligence (AI) startup born out of its sister company Hero Sports, to be its “preferred platform for automated sports and gambling content.” The AP announced their partnership after Data Skrive demonstrated its expertise in generating MLB, NBA and NHL reports.
Data Skrive (the latter term meaning “to write” in Scandinavian) is the brainchild and creation of CEO Brad Weitz and co-founder Jordan Nilsen, the company’s CMO.
Upon becoming CEO, Weitz quickly realized historical methods for scaling content production were not viable solutions for covering a myriad of topics. Without the ability to quickly scale, companies are forced to slow growth, which allows competitors to gain ground and threaten their position in the marketplace.
Their goal isn’t to destroy sports bureaus everywhere; their goal is to make journalism, specifically local journalism, a growth business and not a loss leader.
“Sports coverage is weighted to the top-notch teams,” Weitz told Observer. “But what about local teams, and the ones that weren’t getting much coverage by reporters? We built the technology to produce ready-to-publish content. Newspapers were trying to cover hundreds of thousands of athletes and thousands of teams, with a declining advertisement cost, and more to cover than ever.”
While local journalism may be declining, local appetite for sports has only grown more voracious. Whether it’s the national media covering every game, or hometown newspapers reporting on each sporting contest, the current flawed model of local journalism means Data Skrive is filling gaps, not jobs.
Rather than spending a ton of time cranking out story after story on local sports games with details that can be mined and packaged together in a story, reporters can now be freed up to get that coach or athlete interview, without worrying about missing a deadline. “There’s more time for writing op-eds and features,” Weitz explained. “AI can’t know certain things, about the challenging childhood the player comes from or the hard-luck stories. You don’t have to stay in the box getting the basic story done. It helps you leverage the skillset of the talent at your disposal.”
Data Skrive provides writers with more time to spend telling compelling stories by using colloquialisms, as well as focusing on marketing, targeting the audience, social media and optimizing for reach. And these technology platforms can incorporate voice, and dynamic imagery, with incredible speed.
“March Madness brackets were released by 3 p.m. PST,” Weitz said about the NCAA Men’s Basketball playoff competition. “By 3:15 p.m., we had a preview for every game and were atop search engines because we were first in bracketology.”
Data Skrive’s website even provides a clue as to how the system works—from data input to platform to article output. Prospective users can even receive a demo for details on how it’s all done.
And it won’t only be the sports media space where you see AI moving in, providing machine content. “You’ll see it in politics, technology, finance, real estate, entertainment,” Nilsen said. “Imagine that we can comb over gigs of data to find the analysis. If a shooting happened today, we could find out in almost no time how this event compares to others this year and historically. That’s instant data, not from hours and hours of researching. Automation also provides more accurate information.”
Policing the Program and Providing the Technology
Speaking of accuracy, I wondered aloud how the company does quality control. Has it adopted the “police patrols” strategy of checking out a small sample of data, or does it wait for complaints, the “fire alarm” approach?
“We focus on four things,” Weitz explained. “First, did we get the data? Did that data match our expectations (for example, did Giancarlo Stanton actually hit three home runs)? Does it follow the proper grammatical makeup (e.g., grammar on Twitter isn’t the same as Associated Press)? And finally, did we get it in a timely manner?”
So where did Data Skrive come from? “We built it for ourselves and for our website HeroSports.com,” Weitz elaborated. “The Associated Press asked us ‘How’d you do this? Would you consider licensing the tech to us?’ Now, we are their preferred partner given our approach and ability to produce content for every marketing channel and multiple formats. We think of journalism, content, marketing and audience altogether.”
“It’s a business,” Weitz added, “but it’s also sports, so it’s a lot of fun!”
By the way, here’s the link for the first source on Hawaii baseball and the second one as well, on UC Davis baseball. Can you figure it out which is the human-written one, and which is the machine-generated version now? If it’s not obvious, then the Data Skrive team can say that they’ve accomplished their mission.
John A. Tures is a professor of political science at LaGrange College in LaGrange, Georgia—read his full bio here.