“A galloping horse that seems to be out of control” was how Advanced Compliance Technology Director Susan Bala described artificial intelligence at Global Gaming Expo on September 30, where AI was a leading topic of discussion.
Bala’s panel held at Venetian Expo Center was slugged “How AI is Transforming Sports Integrity and Match Fixing.” Her ACT company was out in force, comprising three of the four panelists. She was joined by Chief of Integrity Kenny White, a veteran oddsmaker and exposer of fixed games, and Strategist Tim Tenpas. Bolt Group founding partner Dean Sloves rounded out the group.
Moderator Bala began by noting that sports betting in the United States had gone from a $6.6 billion business in 2018 to $168 billion size today. Globally, it is $5 trillion, once legal and unregulated (and black-market) books are taken into consideration. Of that, 95 percent is wagered online.
“It’s explosive growth,” Bala bemoaned, “and there are little to no technological tools” for overseeing it. ACT, she said, has compliance silos that monitor 700 sportsbooks and 10 prediction markets, looking for suspect activity. “We have to watch these betting habits in a way we never had before.”
White offered a bit of personal history on match fixing, noting that when he was making book, he was approached by clients who were getting beaten up on Mid-America Conference football action. “There’s no way you should be losing,” White agreed.
He began studying team records and found the University of Toledo to be 28-1 against the spread whenever heavy money was being wagered. He took his findings to Nevada regulators, who couldn’t turn up anything unusual. The FBI did, however, and White’s reputation as a fix-finder was made.
Tenpas, a sharp bettor, described himself as a 1,000-to-one long shot to be on a compliance panel. After all, he organizes syndicates to beat the books at sports wagers, gleaning as much relevant information as he can from informed sources.
“Everything comes down to information,” Tenpas summarized and AI is processing information on a mind-boggling scale now. The opportunities presented to regulators, he said, is far beyond what bettors like he can do. There is, he concluded “a complete sea change” brought on to sports betting by AI.
Match-fixing determinations, Sloves said, can now be done at vast scale and speed, flagging suspicious data in what he called a cat-and-mouse game. But bad actors are adapting just as rapidly.
Humans, he continued, simply can’t compete with the speed of contemporary betting, especially in a climate of one-pitch micro-bets. “It allows a new level of observability,” Sloves said of unsleeping AI.
Sloves also noted that players will become more sophisticated, using AI agents to place their wagers, sometimes seeding bets in suspicious ways, possibly unauthorized ones.
To this, White added that one game could have 300 relevant wagering propositions. But AI can and does enable tracking of whether the athletes are underperforming or not. He said he hoped to get it across to such players that they’re risking everything if they collude, noting the Toledo players threw away their careers for $10,000 each.
However, AI isn’t seen to be a panacea. Sloves observed a large problem with the siloing of data between AI departments. “They don’t really talk to each other to have the right context. If you could have it all in one place, that’s context.” Another problem he noted is that “the world is kind of weary” of data, being subjected to hundreds of alerts per day.
“Knowing the cat is always around the corner,” Tenpas remarked, there’s less incentive now for bettors to evolve to outwit the technology. His kind of betting can’t be done at scale, he said, without the movement of vast amounts of money quickly. “I have people on every continent except Africa, making wagers on every sport.”
Tenpas admitted he used to pump the money lines by flooding books with bets. Now, as lines become sharper and anticipate movement, it’s harder to do.
“The gaming industry might be one of the few, maybe the only one, where the human is paramount,” Bala opined, likening present-day AI to the early days of the Nevada Gaming Control Board. White followed by allowing that he fights with his AI two to three times a day and that it’s crucial to feed it the right data when building it.
Human frailty also has to be taken into account when scrutinizing athletic underperformance, White cautioned. To do his job, he programs the AI to work through insider information and the movement of money, correlating it to sundry other variables: “You’re always looking for an anomaly between the money and the score.”
Potential match-fixers, though, can be tracked through everything from social media to geolocation, as well as the timing of their bets. White didn’t disagree that this had some Big Brother implications and neither did Tenpas. The latter said that “fish bettors” sustain the market, but don’t impact books’ bottom lines, whereas “me and everyone who does I what I do” easily can.
At the end, the question of prediction markets was raised. White was dismissive, calling it “the same as sports betting. It’s just changed the vocabulary.”
He also lamented that yet more match-fixing oversight is needed, even 107 years after the Black Sox scandal. “It all takes time,” he sighed.
