In April 2026, KPMG collaborated with the University of Nevada, Las Vegas (UNLV)’s International Gaming Institute, as well as its AI Research Hub, to release The State of AI in Gaming 2026 report. The report tracks how artificial intelligence continues to shape the industry – and we asked Rick Arpin all about it. 

What were the most notable or insightful findings for KPMG within the State of AI report? 

From our firm’s perspective, we were very interested to see what operators and suppliers said about where they were in adoption. What we found was certainly an eagerness to leverage AI and some decent adoption from a trial-and-error perspective. But we also found the industry is behind other(s) in their adoption, in particular the land-based operators, who were behind the suppliers and online operators in terms of how much they’ve adopted AI. We can talk about why that might be, what we saw in the study and what we know about the background conditions of that, but those were certainly the first ones that jumped out.

In addition, there is a level of disconnect between how regulators are feeling about AI, what they think operators are doing, what operators are capable of doing and where operators currently are. We saw that as an area where there’s opportunity for more communication, education around how AI can be used and getting regulators to at least know what operators are doing.

Touching on that more deeply, where can operators, suppliers and regulators improve in AI infrastructure and expertise?

There’s not enough focus on the governance of AI and how to implement AI responsibly. Now, the good news is there’s time to do that because they’re a little bit behind in terms of adoption compared to other industries, so it’s not as if they’re way down the road and haven’t thought about any guardrails. It’s not a drastic situation, but it’s really important to be thinking about the infrastructure here. In this case, who can use the AI? How do we control the models? How do we review the models? How do we protect our data? All those responsible AI notions are important, and we’re not seeing enough emphasis on that.

While the industry is, again, in the trial-and-error phase of starting to leverage AI, they’re mostly focused on the large-language models and the generative AI use cases. We’re already seeing other industries go far beyond that into the more agentic AI, where it’s processing entire strings of transactions and executing decision-making. The more you do that, the more those responsible AI practices are even more important because our firm’s view is there should always be a human in the loop. If the model is tested and if the transactions are coming out right, we want to gain the efficiencies and the effectiveness of having agents doing things, but we need those controls if you will.

Given the report stated operators and regulators currently “disagree significantly” on how AI is being deployed across gaming, what steps are needed to find a middle ground? 

It would seem we need more open dialogue. Maybe the sessions we’re all having at conferences just aren’t deep enough into specific examples. Part of the issue is operators and suppliers are leveraging AI in areas where they’re not running across the actual regulations they’re subject to. They’re looking in other areas of the business in part because they understand if they go into a gaming area, then they may have to then talk to the regulator, get a waiver, have a regulation change, etc. That takes time.

Why don’t I trial and error over in the hotel department or in a finance area where I then don’t have to worry about the regulations? That’s fine, but then regulators are not getting visibility into how you’re using AI, so somewhere in there is seemingly an answer. I don’t think it has to be terribly invasive. It’s more of an upping the level of collaboration and communication.