Few technologies have pervaded the social conscience and disrupted industries as quickly as artificial intelligence, or AI, has. Far from a buzzword or a flash in the pan, AI has spread its tendrils to reach every nook and cranny of the tech world, including and perhaps especially gambling.
AI’s uniquely widespread nature makes it particularly tough to get right, especially from an ethical perspective. Who decides the right and wrong ways to deploy AI? What requirements should operators and businesses meet to reassure their user base that AI tools are being used for helpful and productive reasons? How—and to what degree—should companies notify their players and customers of AI usage?
Big questions, to be sure. Dr. Kasra Ghaharian, director of research for the International Gaming Institute at University of Nevada Las Vegas, is researching ethical AI usage, including what it means and how businesses can do it the right way.
Ghaharian landed at UNLV to work on his PhD after doing stints in marketing at Las Vegas Sands (LVS) and Caesars. At LVS, Ghaharian was a marketing analyst and product manager for the digital marketing and ecommerce department. At Caesars, he worked in digital marketing. He previously got a masters at UNLV and wrote a thesis about using mathematical approaches for slot-floor optimization.
“I concentrated on leveraging data science and machine learning for consumer protection and responsible gambling,” he said of his previous experience.
His work snowballed into a larger focus on AI. He wanted to explore the “other purposes” for which AI was used beyond responsible gambling. “I became interested in ethical AI use,” he said. “Are the stakeholders who are using AI doing it in a responsible manner?”
Ethics in general an issue for the ages
Ghaharian jumped down the ethical rabbit-hole and had to define, as best he could, what ethics entail. Naturally, it’s a philosophical conundrum passed down and iterated on for millennia, but modern experts seem to agree there are three areas of interest when it comes to ethics of any kind.
“There are three ethical theories that dominate the discourse,” he explained. “One is utilitarianism, in which something is most often deemed ethical if it provides benefit for the majority of people.”
Some might call this the “greater good” model. Perhaps a user agrees to provide their data for AI modeling and it improves a live chat feature that later helps thousands of people. One person’s information being used could be seen as ethical in such a case.
“The second area is called deontology ethics,” Ghaharian said. Deontology comes from the Greek “deon,” meaning duty or obligation. “This branch of ethics aligns with duty. Doctors have an obligation to tell you the truth about your medical condition, for example.”
Ghaharian noted this area of ethics is harder to pin down in gambling. The framework defining ethical practices might be an internal policy. Perhaps a perfect-world scenario is a set of ethical regulations from legislators surrounding AI, but that comes with inherent risks, too, like legislators or people with influence not understanding what they’re regulating (in one of Ghaharian’s anecdotes, he recalled hearing a professional hearing the term ‘algorithmic bias’ and thinking it was a medical issue).
“The third type is virtue ethics, which is really murky,” said Ghaharian. “Virtue ethics might not help the greater good, but they come from a virtuous place.”
Because “virtuous” is just as difficult to define as “ethical,” it’s hard to view ethical AI use through this lens. Businesses often deal with big customer bases anyway, so the first two perspectives provide ample opportunity to brainstorm.
Defining ‘ethical’ use of AI
The unfortunate reality of current AI deployments in the gambling industry is that they’re fast and loose with few clear-cut regulations. There isn’t any single party to blame, and the move to ethical AI use must be collective.
The first step is identifying ethical and unethical uses of AI. In 2023, Ghaharian kicked off a research project that involved interviews with more than 30 industry experts in various roles.
“What these interviews identified,” he said, “is a balance of using AI for ‘good’ and understanding what risks it poses. Obviously, you can use AI for good in the responsible gambling arena.”
Indeed, there are companies focused on identifying problematic gambling behavior using AI models, such as Mindway AI.
Of course, the flip side is the use of AI for targeted marketing or game development. If businesses use these processes, Ghaharian encouraged asking key questions upfront: “Do you have a human in the loop? How transparent are you? Are you taking data privacy issues into account?”
Using information gathered via AI for game development or targeted advertising can have myriad implications. The question isn’t so much what those are, but rather, how best to manage AI use.
Ghaharian is a researcher at heart. He deploys the scientific method and releases his findings. His experience with AI and the people using it has yielded some simple ideas that most companies could implement to protect themselves and end users from unethical AI usage.
First, he referenced the European Union’s data privacy laws as a template for ethical AI implementations.
“Users of a platform should be informed of what data will be pulled and how it will be used by AI,” he said. This information should be easy to understand for the everyday person. “Tech literacy is crucial,” he said. It might start with a simple checkbox acknowledging awareness use, similar to accepting terms and conditions for any online businesses. But it should extend to include clear and concise explanations of how a consumer’s data is being used.
The more eyes the better
Second, Ghaharian warned against “ethics washing.”
It’s similar to green-washing, when a company makes efforts to appear eco-friendly but isn’t actually doing much behind the scenes. In other words, those using AI should take it seriously and do the heavy lifting to establish an ethical framework and stick to it. Ghaharian’s third idea stems from the second.
“Have a board,” he said. “A group of individuals who review and give the go-ahead on AI projects. As a researcher, every project I do, I have to submit to our ethical review board. They say whether I can do it, or they give me changes I need to make to my study design. I think the same should happen with marketing campaigns that use AI.”
An extension of this idea is the rigorous documentation of AI features. Ghaharian continued: “If regulators ask how a company is using AI, that business should be able to pull up a registry of all their use cases. This will make it easier to govern long-term.”
Finally, Ghaharian thinks the legislative pipeline is moving, and he hopes to see AI laws modeled around existing data-protection regulations. Time will tell, but all of his expertise points to the need for strong and decisive action on ethical AI use.
