Mitchell Forster is using agentic artificial intelligence to manage his fantasy NBA team this year.Nicole Osborne/The Globe and Mail
Artificial intelligence is muscling into fantasy sports leagues. It’s making draft selections. It’s setting starting lineups and grabbing guys off waivers. And its value to a fantasy general manager reflects a broader divide between AI superusers and traditionalists.
That divide could not have been clearer last month at the draft for one long-running league. A dozen members of Fun Time Fantasy Basketball – they picked the name back in high school, 26 years ago – sat around a boardroom table in a midtown Toronto office, beer and Halloween candy spread out between them.
At one end, Mitchell Forster had his laptop open to the custom command centre he’d spent weeks finessing in ChatGPT, which provided a dizzying array of projections and game plans that adjusted in real time as the draft went on. At the opposite end, Harley Hayman rifled through three sheets of paper. He had printed out last season’s stats for the available NBA players. Once they were taken, he used a highlighter to cross off their names.
The algorithm-versus-analogue showdown at Fun Time Fantasy Basketball has been brewing for the better part of a decade.
“I’d always built sophisticated Excel spreadsheets to prepare for the draft,” Mr. Forster said.
Meanwhile, Mr. Hayman declared, “I’ve been getting dissed for eight to 10 years for not even bringing a laptop.”
Harley Hayman in the boardroom where he participates in fantasy sport drafts with his friends, in Toronto. He prefers to stick with pen and paper.Sammy Kogan/The Globe and Mail
This is the first time, however, that Mr. Forster is relying on agentic AI to manage his team. It was a natural evolution. “With the AI boom we’re now in,” he said, “I can tell ChatGPT, ‘Create me a tool as if you were the best fantasy basketball expert on Earth.’”
ChatGPT whipped up a strategy for his season, assembling a roster that prioritized meaningful assists and successful free throws while supporting field goal percentage and blocks. It also assigned each player a value for those categories, then piped in with advice throughout the draft. “[Milwaukee Bucks centre Myles] Turner supplies blocks and threes, but his projected FG% does not balance Trae,” it cautioned. (Like most NBA announcers, ChatGPT is on a first-name basis with Washington Wizards guard Trae Young.)
That’s a very small sample of what it’s come back with. “This is definitely overkill,” Mr. Forster acknowledged, adding that he’ll probably dial down the model next year.
AI isn’t calling all of the shots. “I’m still in the loop,” he said. “I still watch games and look for stars in the making. I don’t think AI can do that yet.”
But by off-loading much of his league’s time-consuming labour – poring over endless fantasy stats; staying on top of what’s happened in the NBA – he frees himself up to focus on other aspects of his life.
How to train your robot
“As we get older, time is ever more precious,” Mr. Forster said. “We all have kids and families and jobs. So if I can save myself some hours by using AI, hey, why not?”
It makes sense he’d see it that way: Mr. Forster works in tech software sales and turns to AI often. Mr. Hayman, on the other hand, isn’t a habitual user of agentic AI. He’s not against it – don’t be fooled by the highlighter and printouts – he just has an inverse take on the free-time issue.
Mr. Hayman with a list of players. ‘Winning takes shrewd pickups and sharp trades that require working the other GMs in the league,’ he says.Sammy Kogan/The Globe and Mail
“I’ve got a wife, two young kids at home and a job with a lot of moving pieces,” said Mr. Hayman, the director of an overnight summer camp. “I do not have time to go down the road of learning these AI programs and having them train me.”
He appreciates that there are benefits to the technology. But it’s more efficient for him to lean on the basketball knowledge he’s acquired over the past quarter century.
“Winning takes shrewd pickups and sharp trades that require working the other GMs in the league,” Mr. Hayman says. “I’m going to use my brain for that.”
It was inevitable that fantasy sports leagues would find themselves weighing the usefulness of AI. Roughly one in 10 Canadians played fantasy sports last year, but nearly 80 per cent of us now use AI tools, according to Leger, up from 47 per cent a year and a half ago. Although Pew Research found the most common application remains basic information searches, through chatbots and Google’s AI Overviews, AI personal agents surged in popularity in the past two months. Meta’s Muse has been the most downloaded app in Canada since its September debut.
AI personal assistants force Canadian companies to weigh customer access against security
Renee Miller, a professor of neuroscience at the University of Rochester and chief executive officer of Franchise League Football, has clocked a rise in the reliance on AI across fantasy sports. Most of the few thousand members of her league are holdouts, she said, preferring to steer the decision-making themselves. But that doesn’t mean they’re wholly opposed to the advantages of these tools.
“My league is actually looking into building an AI bot that would step in for absentee managers,” Prof. Miller says. The agent could set fantasy lineups, make adjustments and even add a few waiver pickups. “It would keep the league competitive, because no one wants to play against a team that only has three active players in their starting lineup.”
A skeptic’s guide to AI
So how big a presence should AI have in a recreational hobby? That’s been the subject of much debate at another Toronto fantasy basketball league, where more than half of the 10 members began using AI for their teams this season. Arguments volleyed between Nick Poissant, a pen-and-paper traditionalist, and the league’s Claude Code power users.
One AI-savvy member built an exhaustive program that analyzes multiple fantasy-stats providers, every relevant news source and the league’s own history to determine the best strategy for a winning team. He just volunteered to share the tool with his fellow managers, positioning it as a way to level the playing field.
Mr. Poissant is unmoved by that offer. “I like trying to figure it out on my own,” he said. The challenge is worth the investment of his time. He’s willing to stick with the league this season, but maybe not much longer. “If it becomes guys just punching buttons on their laptop, I’ll probably move on.”
Over at Fun Time Fantasy Basketball, Mr. Hayman is happy to be an AI resister among the GMs in his league. “It’s honestly a badge of honour for me. I’m going to keep competing. I’m going to keep winning.”
And though it’s early in the season, he’s quick to point out that he’s second in the standings – one spot ahead of Mr. Forster.