Artificial Intelligence (AI) has influenced how the gambling industry operates in multiple ways. It has widened the industry’s horizons, creating an environment where machine-operated capabilities are almost limitless.
Practically every area of the industry takes advantage of the applications of AI, from developers creating games to operators monitoring customer play.
AI is also changing how the world develops outside the industry, becoming a key component of self-driving cars and medical advancements.
Last year, Grand View Research estimated the global AI market was valued at $62.35bn in 2020, and predicted that this would rise to more than $930bn by 2028.
Reflecting on recent years, Paolo Personeni, managing director of managed betting services at Sportradar, sees AI as indicative of how the industry has matured and essential to its ongoing development.
“Our industry has undergone significant digital transformation in the past five years and artificial intelligence has been critical in its evolution,” he says. “[It] is an essential component of product development.”
It could be argued that the gambling industry is an ideal testing ground for AI. AI functions through harvesting data, which allows it to analyse statistics and occurrences to continually improve upon its accuracy.
And data is at the heart of how AI operates in the gambling industry, as Lloyd Danzig, founder and managing partner at Sharp Alpha Advisors, points out. “The gambling industry relies heavily on the leveraging of historical data to anticipate future outcomes and behaviours,” says Danzig.
“This is a core competency required for game design, oddsmaking, risk management, customer profiling, rewards programme optimisation and fraud detection.”
Aiming for the goal
AI’s use in sports betting is particularly key, says Personeni, as an asset to ensure Sportradar’s products and solutions are as accurate as possible. “For Sportradar, one of our core focus areas is on utilising machine and deep learning to develop exceptional computer vision capabilities, which is a form of artificial intelligence that teaches models to interpret and understand videos and images,” he explains.
“While we, as individuals, are very good at understanding what is happening in front of us at sporting events, we have limited capacity in the amount of data we can manually interpret, record digitally and upload.”
This can lead to AI picking up on specific aspects of gameplay, which Personeni sees as a bonus for customers.
“Computer vision provides a hundred-fold increase in the level of statistics and data that we can produce and then feed into our products and services that we offer to our clients,” he says.
“For example, by calculating velocities and trajectories of players we can predict when players will go offside or predict when an attack is mounting.”
AI that goes further than data crunching, focusing on to-the-minute information, is already being used in sports betting and esports betting.
Real-time AI is the top choice for these areas, wherein data providers must provide the most up-to-date information on gameplay. Pandascore CEO Flavien Guillocheau revealed last year that the operator creates 300 data points every half-second for League of Legends with its AI, with an accuracy of over 99%.
Not only can AI facilitate betting by making the process much smoother, but it can also act as a safeguard and create areas of engagement.
AI can be designed to pick up on any and every aspect of a game, which lends to its boundlessness. In 2019, hardware intelligence company MicroTeam launched a pair of AI football boots which allow gameplay data to be captured within a millimeter of accuracy at high speed, including attack routes and distance travelled without the ball. This is then analysed and used for player and match assessments and sports betting platforms.
Similarly, AI is used to analyse and predict gameplay in esports to provide the best experience for bettors. Esports Technologies is one of the latest companies to implement AI in this way, filing a patent application last October to cover AI that can create odds models for esports betting tournaments.
And just as AI facilitates new permutations of gambling products, so it can act as a safeguard for the adverse effects of the industry. Leigh Nissim, CEO of Future Anthem, sees AI as a road to reducing problem gambling. However, he adds that the extensive types of technology make data collection challenging, which could stifle the full effectiveness of AI.
“In responsible gambling, there are quite traditional practices of how to prevent player risk,” says Nissim. “You’re modelling on other players who have cancelled and self-excluded. As consumers move to other tools, such as open banking and other ways of depositing, it’s increasingly hard to see how they are depositing. That data set is quite challenging.”
AI features heavily in customer interaction, particularly in maximising the player experience in areas such as customer service. Danzig emphasises that it is also of great use in detecting and preventing problem gambling.
Nissim questions whether operators are making the most out of AI in terms of customer service.
“I think people are making strides in trying to prevent player exits or departure, or trying to optimise their marketing funnel,” he says.
“There’s increasing sophistication in the use of complex models that predict those characteristics. But are they using AI in real time to improve consumer experiences?”
“Commercial AI tools are extremely capable in cases that rely on pattern recognition across large data sets,” he adds. “These capabilities allow for the efficient analysis of historical user behaviour and engagement to determine how to maximise the experience for a particular user or when to flag behaviour as being potentially harmful.”
As the gambling industry advances, there are ongoing conversations on how to optimise, enhance and expand the sector further.
While much of these conversations focus on digital applications, AI could play an equally important role in the land-based sector. According to AI cloud service DataRobot, AI can be implemented on casino floors. The service can be used to predict the level of play that could occur on the casino floor and adjust staffing levels to match.
The rest of the industry is quickly following suit. However, AI should not be seen as a remedy for all sector shortcomings. Danzig explains how the initial implementation of AI can come as a shock to businesses, particularly if they have a cavalier attitude.
“When businesses first embark on the journey to integrate AI solutions, they often find that the process will be much more costly and resource-intensive than originally expected,” he says. “In part, this is due to a misconception that AI is a one-size-fits-all panacea that immediately improves efficiency and profitability.”
According to Danzig this is not a problem-free solution and relies on a certain standard of data to work efficiently.
“Once implemented, a primary downside to relying on AI solutions is that they can often exhibit unexpected behaviours that result in quickly compounding issues given the speed and scale with which they are deployed,” he continues. “AI solutions are only as good as the quality of the data being used as inputs, hence the phrase ‘garbage in, garbage out’.”
A longstanding concern with AI, which has been documented thoroughly in popular media, is its potential to dominate society. However, this is not an industry concern.
Nissim believes that while AI could lead to reduced employment demand for certain manual tasks, many roles will be needed in the data side of AI.
“I don’t think [AI] should threaten jobs,” says Nissim. “I think what it should threaten are manual processes… And actually, that’s where you find mistakes arise.”
This, he says, will lead to a growth in opportunities and employment.
“Maybe manual processes will reduce, but what you’ll see is technical upweighting. There will be more data scientists, more data engineers, more technicians and more analysis.
“It’s not about job threats at all. The focus should be betterment and improvement – delivering innovation to provide an improved experience for the player and to help protect operators.”
Danzig agrees with this, emphasising the need for human labour in keeping AI running smoothly.
“While it is possible that some jobs in the gambling industry will be lost to automation, AI solutions today rely heavily on human input and maintenance,” he says.
“In the next five years, the implementation of AI and machine learning tools across the gambling industry will create a significant number of new jobs and also enable stakeholders to refocus efforts toward more productive, and ultimately job-producing, functions.”
According to Personeni, Sportradar’s AI-based capabilities already require a workforce to maintain, with the company employing “more than 700 engineers and 40+ AI and machine learning experts to create 20 billion data files ingested from inception, 30 million odds per minute and near 100% total accuracy.”
While AI has been revolutionary for the industry, there is natural curiosity towards where it could go next.
It’s difficult to say, due to a combination of how boundless AI is proving itself to be and how fast the gambling industry is evolving. As Nissim adds, “In some instances, [AI] can only develop as fast as the technology around it.”
If the last few years of technological development are indicative of the future, AI will able to keep up with and, at times, exceed what the gambling industry can offer.