Artificial Intelligence (AI) is becoming increasingly integrated into online gambling platforms. While it offers the potential for early detection of risky behaviors, concerns arise that the same technology might exacerbate addiction if not properly monitored.
Emerging multimodal AI models, such as OpenAI’s ChatGPT, are being recognized for their capability to serve as a protective measure. However, experts warn that these innovative tools could also lead to more harm if used improperly.
Dr. Kasra Ghaharian, research director at the University of Nevada’s International Gaming Institute in Las Vegas, emphasized the heightened risks associated with easier access to gambling activities. “You no longer have to get in your car to drive to a casino; you just sit on your couch and download an app,” Ghaharian stated. “I can’t overlook the potentially more engaging or addictive nature of this product.”
Studies show that the increase in online gambling access has led to greater exposure and intensity, particularly among younger individuals. This shift has enabled AI models to detect risky behavior more effectively, while also personalizing engagement, potentially leading to greater risk.
As gambling becomes more entrenched in digital ecosystems, AI’s role evolves beyond a mere personal assistant function—like querying ChatGPT about betting options—to being embedded in gambling platforms that monitor user behavior. However, this same technology can also inadvertently deepen an individual’s addiction.
AI Models for Identifying Risky Gambling Behavior
Online gambling platforms continuously generate vast amounts of behavioral data, such as bets placed, deposits made, and transactions canceled. Multimodal AI models analyze these diverse signals, allowing platforms to comprehend not only what users do but how their behaviors evolve over time.
Dr. Zahra Shakeri, an assistant public health professor at the University of Toronto, discussed how this transformation has changed the landscape of gambling risk and its detection. “Traditional methods observe the bets, but multimodal approaches capture the narrative surrounding those bets, highlighting early signs of harm,” she remarked.
Shakeri noted that conventional safeguards often rely on outdated standards, like spending limits or session lengths, which struggle to adapt in fast-paced, AI-enhanced gambling environments where options and prompts are constantly evolving.
With multimodal AI models, interventions for risky gambling, such as suspending a player’s account, can occur before a casual wager escalates into problematic behavior. “If we limit ourselves to traditional methods, we catch harm too late,” Shakeri explained. “It’s like diagnosing lung disease solely by counting cough drop purchases. Multimodal approaches examine data from multiple sources, offering a clearer context for identifying issues.”
The Risks of AI Models in Gambling
Experts warn that the same AI systems capable of recognizing vulnerabilities are often optimized for user engagement. Consequently, both risk identification and risk creation can arise from the same multimodal systems.
“The application of AI in this domain is a double-edged sword,” Shakeri noted. “Clear policy regulations should target both the use of AI and its potential to help address associated problems.”
As someone involved in research on multimodal systems, Shakeri mentioned that models designed to detect risky behaviors could also be adapted to identify users most susceptible to incentives or promotions, potentially nudging vulnerable players toward deeper engagement.
The increasing use of large language AI models presents further challenges, particularly as players consult these tools for guidance in their betting decisions. Ghaharian predicts a future where individuals may leverage AI to enhance their gambling strategies, despite these tools being intended as safeguards.
“We will undoubtedly see many AI-based betting assistance products,” Ghaharian said. “Regulators need to communicate how these technologies function so users are aware of the risks and boundaries, utilizing them safely and responsibly.”
If a gambling app features an AI chatbot that intervenes during risky behavior, players might also interact with the bot to inquire about wagers or strategies, complicating the understanding of interaction boundaries.
Current AI Usage in Online Gambling
As gambling companies begin to embrace the dual potential of multimodal AI systems for both risk assessment and intervention, they are changing the timing and nature of interventions.
Fanatics Sportsbook has launched AI-driven tools as part of its responsible gaming framework, complementing existing safeguards. Anthony D’Angelo, head of responsible gaming at Fanatics, confirmed that their practices align with New Jersey’s responsible gaming standards, even in jurisdictions lacking specific mandates.
They have incorporated an AI-based risk-scoring system, Neccton, created by gambling researcher Dr. Michael Auer. This system evaluates numerous behavioral indicators, assigning a risk score to players so responsible gaming teams can prioritize their reviews and intervene more effectively.
“Neccton follows a proprietary algorithm that we don’t alter … based on Dr. Auer’s markers of harm,” D’Angelo explained. “For instance, if a player frequently cancels withdrawals or struggles with insufficient funds, their score rises, indicating that intervention may be necessary.”
Instead of merely reacting to singular triggers like significant deposits or a flurry of bets, the system focuses on players whose behaviors show sustained or rapid changes, allowing for earlier intervention in their betting patterns. “We implement measures, whether it’s communication or deeper review by the responsible gaming operations team, to decide on possible suspensions,” D’Angelo noted. “Depending on the situation, a player could face a temporary or permanent suspension.”
D’Angelo clarified that the AI system does not diagnose addiction or operate independently. Instead, it standardizes evaluations, minimizing subjectivity and helping human teams consistently recognize concerning trends.
Furthermore, large language models are employed internally to analyze customer communications for signs of distress, prompting referrals to responsible gaming specialists.
Policy Decisions Regarding AI’s Impact
As AI becomes more intertwined with gambling platforms, the ultimate influence on gambling behavior will heavily rely on regulatory frameworks. Ghaharian anticipates that more AI-driven betting assistance tools will likely emerge, increasingly obscuring the distinction between guidance and influence.
“Many gambling operators may not fully grasp how these technologies are utilized and the associated risks,” Ghaharian explained. “This presents an opportunity for regulators to mandate safety evaluations for large language models, especially those engaged with customers.”
AI WON’T CAUSE A SUDDEN JOB APOCALYPSE, BUT A WAR OF ATTRITION
Shakeri concluded that the vital question for policymakers is not whether AI will influence gambling behavior, but how it will do so. “The focus should not just be on the adoption of AI,” she said. “It’s about what objectives we allow it to optimize.”
