Paula Murphy: The integral role of AI in the player protection revolution

Everywhere we go these days, everyone is talking about AI. There are concerns about ethical implications and perceived risks. Stories of students cheating on their term papers, humans being replaced by robots and a collective fascination with ChatGPT. Earlier this year a group of tech experts and leaders, including the likes of Elon Musk and Steve Wozniak, co-founder of Apple, signed an open letter calling for development to pause, citing "loss of control of our civilisation". The mainstream media are in a frenzy and in cafés and bars all over it is the hot topic of conversation.

This is however only one side of the story. Yes, AI can be, and is, developed by bad actors just as easily as those who use it for progress. Whatever your opinion though, AI is here to stay. It is all around us in our day-to-day lives from our Spotify daily mix to our social media feeds and object detection helping us avoid accidents in our cars. In the medical world, AI is helping save and enhance lives by enabling faster detection and diagnosis of serious illnesses and assisting in the development of new drugs. In business, we are harnessing the power of AI to do the heavy lifting and streamlining of processes.

The idea of putting the genie back in the bottle doesn’t seem that realistic. Despite the concerns we hear about AI, I don’t believe the answer is to pause all of that positive development. Far from being afraid of new technologies, should we not harness their power and strive to understand the possibilities they afford us if deployed effectively? Nowhere is this truer than the world of player protection and the identification of potentially problematic play in igaming.

The importance of player protection has become more and more central to the running of a successful gambling business in recent years. Regulators all over the world are expecting operators to be able to identify players who are at risk of experiencing gambling-related harms and deploy effective intervention strategies. In addition, operators are waking up to the benefits associated with understanding the risk profile of the individual players who make up their customer databases in terms of maintaining a sustainable player base and avoiding huge regulatory penalties. Customers, especially the younger generation, expect operators to have a social conscience and will vote with their feet if they don’t believe the values of the business match up to their own.

In addition to this, AI has also significantly improved customer support interactions within the industry. Operators now deploy AI–powered chatbots that can provide customer support 24/7, providing immediate assistance to players by addressing regular queries and signposting them toward other tools and services. This round-the-clock support ensures that players receive aid regardless of the time of day.

There is obviously a huge value chain involved in protecting your players from providing the right training for your employees to understanding the nature of problem gambling and creating an environment of positive play for your customers. It all starts with accurate detection though. Scanning a database of thousands or in some cases, millions of individuals looking for signs of potential vulnerabilities or harms in real time, is not something that is scalable without the use of technology.

Furthermore, player protection has evolved, it is true that very high-risk customers account for only a small proportion of all gamblers, but the goal needs to be early identification of potential problems coupled with solid intervention strategies to avoid that kind of escalation in the first place or where necessary deploy controls on accounts and point players in the direction of help. In short, protecting all of the players all of the time.

Doing this accurately is where the value of AI comes into its own. The key to accurate detection is the proxy and understanding of what you are looking for in that sea of customers. Utilising a data-driven approach to understand each player’s individual gambling behaviour is how we do this at Mindway AI. We bring together human intelligence in the form of leading experts in psychology and problem gambling and AI technology and train the algorithm to replicate complex human decision-making based on the patterns it observes in the spin-level data.

Markers of harm are important, thresholds are helpful but in isolation, they only tell part of the story and there is a danger of high numbers of false negatives or false positives. Only by assessing the behaviour in this contextual way can you get a clear and accurate view of what is really happening. The end result is akin to having an army of psychologists monitoring your players around the clock.

Good AI solutions also need to be explainable and understandable. Once we understand exactly why the system is flagging a behaviour as problematic, we can communicate with the customer in a meaningful way. Personalising the intervention so that the customer identifies with the conversation and offering safer gambling advice that is helpful and relevant to them. Such interventions can be both automated or manual depending on the severity of the case or even the profile of the player.

Imagine all the communication tools we have available to us and then imagine linking them up to a powerful analytics system that will drive these tailored interactions based on customer behaviour. In terms of providing a personalised customer experience, the integration of AI algorithms into online casino platforms facilitates this as well as optimizing operations. AI–powered recommendation systems that can answer certain queries at unsocial times of the evening, this also improve customer engagement and satisfaction, leading to increased customer retention and loyalty.

As new studies are published and our understanding of gambling behaviour evolves, so does the player tracking, because it replicates the decision-making of the human expert panel. Cultural and generational differences in behaviour can be accounted for as can different regulatory requirements across multiple jurisdictions or unique behaviours associated with different gambling verticals. Anything a human expert can understand and factor into their decision-making, the algorithm can be trained to replicate on scale.

Our RG teams become more effective because they can prioritise the customers who really need their attention. They also enter such conversations and account reviews armed with much more insight and intelligence about what is happening with each individual customer. Decision-making about safer gambling strategies and controls becomes more informed and targeted to each individual player's circumstances.

A data-driven approach to player protection that combines human expertise and artificial intelligence resulting in earlier identification of potential problems and more informed interventions will enable us to steer more players back to a sustainable level of play rather than being lost to problem gambling. It is not about letting the machines take over. It is about identifying the customers who are experiencing gambling-related harms and helping them whilst allowing others the freedom to gamble safely for entertainment.

Image credit: Casino Guru News