Safer Gambling Experts Warn of Losing Human Connection in Race to Embrace AI

^{Panelists Pedro Romero, Dr. Melissa Salmon, Christina Savvouri, and Ryan McCarthy at New Horizons during a discussion at the Safer Gambling Conference in Vancouver. (Photo: courtesy of British Columbia Lottery Corporation)}

It’s becoming increasingly common for casino safer-gambling programs to be supercharged by AI. But at the New Horizons in Safer Gambling conference in Vancouver last week, researchers and industry experts warned that the pursuit of increasingly refined algorithms is putting gamblers at risk by neglecting the human factor.

AI is seeping into every facet of modern-day life, and the player protection teams at gambling operators are no exception. Tools like Mindway AI’s Gamesense system promise to help spot at-risk customers before they begin gambling recklessly, and operators have been quick to adopt them. Australian casino giant Crown Resorts celebrated the power of its new partnership with the platform after integrating the tech last year.

“This will greatly assist in identifying risks and preventing harms from occurring in the first place,” said the company’s head of responsible gambling, Dr. Jamie Wiebe.

Spotting Risk Is the Easy Part

But even if safer gambling teams have gained the Minority Report-like ability to spot harm before it happens, experts have warned that this is really only half of the battle. How staff speak to customers in a way that convinces them to modify their behaviour is an arguably more challenging skill, they say.

“We have become very good at identifying people [who have] risk,” said Pedro Romero, the chief of safer gambling partnerships at Betblocker.

Speaking at the New Horizons in Safer Gambling conference last week in Vancouver, he said that the industry as a whole often lacks the skills to interact with gamblers in a meaningful way.

Telling someone at risk of harm to simply take a break tends to backfire, Romero cautioned. “The likelihood is they’re just going to take a break from your company” and gamble somewhere else instead.

Why One Warning Isn’t Enough

Speakers at the event also warned that even AI-backed certainty that someone is gambling recklessly only goes so far. Getting that person to actually change course takes repeated, sustained effort, not a single flagged alert.

“One interaction isn’t going to save the day. One interaction has absolutely no impact,” said Dr. Melissa Salmon, the senior research lead at gambling-focused researchers Focal.

She argued that it’s the accumulation of “multiple interactions” that works, not any single intervention.

Christina Savvouri of Mindway AI warned that failing to make a real dent in player harm will eventually doom the industry to tougher regulation. “It’s about changing the narrative and general culture of our industry and making it unacceptable to not care about your players and your client base,” she said. “If not, the corrections will take place naturally” — pointing to tax hikes and fines already landing on operators elsewhere.

Gamblers Are Already Outsmarting the Algorithms

A better grasp of the human factor matters beyond just difficult conversations — it also matters because tracking systems themselves change the behavior they’re meant to catch. Salmon revealed that her research shows human ingenuity, particularly in addiction, should not be underestimated.

She pointed to findings from Australia showing that when slot players learn they are being tracked by AI, they deliberately shift their play patterns to avoid triggering safer-gambling algorithms. Salmon cautioned floor staff not to ignore potential interactions with their guests just because those interactions haven’t triggered an algorithm's safeguards.

“Models can rule people in, but they can’t necessarily rule people out,” Salmon said, adding that human intuition in regular friendly interactions must remain a part of a casino’s strategy to protect its customers.

Romero, whose employer offers an app that allows individuals to block themselves from accessing gambling websites, said he had witnessed the lengths that people will go to circumvent safer gambling tools, even when they themselves have chosen to use them.

The Players Falling Through the Cracks

Later in the conference, Salmon presented work by Focal's research team showing that commonly used affordability models fail to detect players who are spending more than they can afford. Operators worldwide are increasingly being required by regulators to assess how much their customers can afford to spend, and to intervene when they suspect players are losing too much.

Focal’s project focussed on gaming machine players and found that, while they fit many traditional stereotypes of a healthy gambler, they are, in fact, digging themselves into a financial hole. Across players studied in the UK, New Zealand and Australia, 4.8% were overspending to the point of needing to borrow money or cut back on essentials to fund their gambling.

Researchers found that these individuals routinely slip through detection algorithms because most models are calibrated on the assumption that problem gamblers will spend multiple hours at the machines.

“Overspending players are playing slower, playing shorter sessions and playing less overall,” said Salmon.

She added that they may even be losing less per-session than players who display other signs of addiction, but this lack of “extreme play” does not mean they are avoiding harm.

Left to their own devices, without interaction and support from casino staff, these players will fall into debt spirals from which they may never fully recover.

Focal says it has developed new tracking techniques backed by monitored machine learning that enable operators to find players gambling money they can’t afford to lose. Among these was the discovery that almost all players who fall into the trap of overspending experience a big win early in their gambling.