Online gaming companies are looking to artificial intelligence to help them take a more personalised approach to preventing players from developing a gambling addiction, with one possibility being individualised messages via chat boxes based on the player's specific play patterns.
As regulators in Europe, North America and other global markets focus on play monitoring and intervention as a key compliance requirement, London-listed Playtech is funding an academic research collaboration to measure the effectiveness of personalised responsible gambling solutions using data and artificial intelligence (AI).
The four-year project, which launched last year, involves a team of psychologists at Erasmus Rotterdam and the University of Amsterdam who are developing a library of messages and responsible gambling tools.
Francesco Rodano, chief policy officer with Playtech, said the psychologists will utilise the risk insights provided by the company’s BetBuddy product to conduct trials with real players in both the Netherlands and the UK.
“This research will further enhance our understanding of player behaviour and enable us to provide more effective personalized interventions,” Rodano told Vixio GamblingCompliance in an interview.
Rodano added that AI can allow an operator to figure out a player’s behavioural profile, which then enables them to take appropriate action based on that information.
“It is indeed a revolutionary project with the potential for significant scientific advancements,” Rodano said. “Through extensive trials, we aim to develop a library that is scientifically proven to be effective.”
In the second phase of the research, psychologists will incorporate artificial intelligence, specifically reinforcement learning, to cautiously enhance these messages. This means that if the system determines that the current messages are not optimal, it will learn and improve them over time.
An additional area being explored, according to Rodano, is “the use of generative AI for player interactions”.
Instead of simply suggesting a break, a chat box can engage in a conversation with the player based on their behavioural profile and propose actions such as setting betting limits or a cooling-off period. Rodano said this approach could be scaled up to handle interactions with multiple players automatically.
“While we are currently investigating this possibility, no concrete developments have been made yet,” he added.
When asked when they expect wide-spread rollout of this product, Rodano said in-play interactions with a bot and messages is already operational, although more AI-driven conversations “may take a couple of years.”
The exploration of AI and machine learning in responsible gambling comes as regulators are increasingly expecting online gambling operators to observe player behaviour and then appropriately intervene based on patterns of play.
Such a requirement is already in place in the UK and Ontario, where operators have been subject to enforcement actions for failing to appropriately monitor or respond to player data. Regulators in various other markets, including Germany and New Jersey, also require the use of monitoring systems for responsible gambling purposes.
In the future, AI could be an alternative to human intervention based on transactional data, with Rodano noting that people are generally becoming more accustomed to interacting with chat boxes in other areas.
“It is worth considering that when individuals have issues with gambling and need to address them, they may hesitate to approach a human due to potential feelings of judgement. This psychological barrier can impact the effectiveness of the conversation,” he said.
“However, there have been experiments conducted with other types of addictions where chat box conversations have proven to be effective in guiding the discussion,” he said. “Therefore, the potential for utilising AI in this context appears promising.”
As part of its responsible gambling strategy, Playtech has leaned on BetBuddy, the responsible gambling analytics platform that it acquired in 2017.
Rodano describes BetBuddy as a behavioural analytics tool that uses AI to analyse the behaviour of every single play in the customer base of a gambling company and tries to match those behaviours with the behaviours of known risky gamblers.
Examples of behavioural markers could be chasing losses, declined deposits, long session times or increased frequency of gambling.
In land-based casinos, staff are trained to directly observe gamblers. If they notice erratic behaviour or any other signs, they can then intervene and try to engage with the player.
Artificial intelligence leveraged by BetBuddy, Mindway AI, Entain's proprietary ARC solution or other technologies can give online gambling operators the opportunity to replicate this form of observation using the vast among of data collected on players’ activities.
“At BetBuddy, we have focused on developing explainable AI,” Rodano said. “This means that instead of just providing a score for each player, we also identify the specific behavioural markers that contribute to that score. This is crucial because it allows us to engage with players in a personalised manner and protect them as a whole.”
For example, if a player is chasing losses or increasing their deposits, rather than merely suggesting the player may have a problem, BetBuddy can suggest setting limits on deposits, taking a break after playing for a long time, or simply asking the player if they are still enjoying gambling.
“This personalised approach has been proven to be up to 20 times more effective in eliciting a positive response from players, as demonstrated by our research,” Rodano said.