Cutting-edge tech developers are making advancements in video and computer-vision technology using artificial intelligence for surveillance that will change how the casino floor is managed and monitored.
That was one of the takeaways from a session at the World Game Protection Conference last week in Las Vegas, where CEOs explained their vision of video surveillance in the future and why the technology will be a game-changer for the casino industry.
The panel included Nicholas Irving, the CEO and co-founder of Eulerion Technologies Inc.; Chencong Ren, the founder of Aeyesky Inc.; and Trevor Outman, the founder and CEO of EagleSight, whose AI technology monitors every camera and sends alerts when it detects cheating attempts, such as past posting, bet capping, and pinching, and catches dealer errors such as pay on push, fail to collect, and paying a loser.
“Imagine our world if computers monitored table games and what would it mean,” said conference host Willy Allison, who moderated the panel discussion. “What if the rating of players was 100% accurate, because the computer gave you this information. What if computers could detect cheating, card counting, and advantage play. What if computers could analyze the game efficiency, speed, dealer proficiency, and errors. Would that be a cool thing?”
Computer vision can make decisions in real time, Allison said. If the surveillance camera is smart, everything is being watched, he added.
“I wrote a white paper a month ago called ‘Seeing Is Believing,’ Allison said. “It is my view of why I think this technology is going to be awesome. It’s going to help us.”
Allison said computer vision is already doing amazing things outside of the casino industry, such as with driverless cars, but he noted one of the challenges in the gaming industry is technology being used to recognize bets. From overhead cameras, it’s difficult to see the amount wagered and he asked how that obstacle could be overcome.
Outman said they recognized a need in the marketplace for accurate player ratings and they got up to 90% accuracy in the lab, but when deployed found challenges.
“I don’t think we are quite there with computer vision to do it accurately where it’s super reliable,” Outman said. “We’re getting close, but we always found problems, like where chips were stacked.”
Outman said that the key is partnering with a company that provides RFID chips, which is expensive and disruptive, but he called it a “foolproof way of getting accurate player ratings.”
Ren said they have been working to overcome the difficulty of reading chips.
Irving said they’re able to detect bets without using RFID technology and noted that the way people should view computer vision and AI as a whole is by looking at autonomous vehicles and military drones. He said his company has taken people from those industries to help develop their technology and help with depth perception.
“We estimate chip stacks and know the amount by using existing cameras, as long as it’s high quality enough,” Irving said of their testing in the lab.
Outman said they’re having a conversation with customers about replacing a strategic one or two overhead cameras with the right dedicated angle.
“Operators are open to placing dedicated cameras and taking advantage of the IT we built to read chip stacks,” Outman said. “It is still fraught with challenges, and we’re not going to be accurate with the challenges of occlusion, lighting, and players playing with chips. It might come to fruition in the next year or two.”
Outman said dealing with casinos can be challenging. Levels of bureaucracy, decision making, and budgets have to be accounted for that create challenges and necessitated a pivot by his company.
Ren said casinos need 100% accuracy and 80% doesn’t cut it.
Outman said they found that among chip manufacturers, the thickness, gloss, and finishes can have a dramatic impact on the accuracy. “It’s not a layup,” Outman said.
Irving said they have 92% to 97% accuracy with chips, while with cards it’s 98%, but it depends on the environment.
“What was unique working with operators and seeing their live operations is that the casino environment is never what you expect,” Irving said. “There are so many things going on, and smoking interferes with camera quality. But the underlying tech has been developed and deployed by companies using cutting-edge computer tech (in other industries). It will come to casinos and help operators make their jobs easier. I am passionate about it, and have engineers who are like minded and think casinos are the perfect place to use this technology from other emerging industries.”
Irving said there’s a problem when it comes to new technology trying to shove itself into an industry that doesn’t understand it. The solution is integrating it properly.
“The one thing I learned is the need to adjust this research mindset to the industry and make it usable for casino executives and their needs,” Irving said.
Outman said his company is narrowly focused on function, reliability, and accuracy for cheating events and dealer errors. Their card accuracy is 99%.
“Our operators want reliability,” Outman said. “They don’t want false positives. We’re laser-focused on that. Secondarily, as we’re talking to them, we’re unpacking all the data being stored in the database from the video. What’s coming from those conversations and what they want are hands and rounds per hour and side-bet utilization ratios. Those are two examples, and there is so much there more, like player behavior.”
There are big differences between what’s desired in Macau and Las Vegas, Irving said. In Macau, operators were focused on collusion rings and how people were finding new ways to cheat and have lost tens of millions by not identifying microtrends.
“This was different from other places like Vegas, where they said high-roller analytics is the math they would like to look at,” Irving said.
Others are interested in combating comp hustlers, who say they have a higher bet amount than they actually wagered, Irving said.
“Technology all of us are developing can be used to perfectly understand that data and create a profile and easily understand and know all of your players’ bets and betting history,” Irving said. “You can see it through the camera system. Using a computer vision system to analyze and record that is a real possibility. That’s different from identifying cheating and harm in casino environments. That’s the dual structure I see in talking to operators, but the industry is constantly changing.
“I think tribal properties will have a different outlook than corporate properties. It really depends on the property and everyone has their own unique needs. But a lot of these problems can be alleviated by using computer vision systems.”
Casinos want a tool that optimizes their labor hours, as staff are manually looking at dealer errors and cheating, but it’s impossible to have a one-to-one coverage ratio, Outman said. The key is that operators want reliability and tools to optimize labor hours and the ability to “put wins in the surveillance team’s bucket” and increase their recovery rate two to four times.
While surveillance staff fear that the technology is coming for their jobs, that’s not the case, Outman said. A human needs to be in the loop to review, validate, and take action on what the technology is pointing out to them.
“If anything, we are going to potentially increase the need for surveillance operators and actually provide more work,” Outman said. “We’re turning surveillance in what’s perceived as a cost center to a profit center. That recovery rate of two to four times doesn’t happen by AI. That AI solution is putting the operator’s eyes on that event in real time, but it’s up to operators to determine what to do.”
Irving said when it comes to AI, the work will change for people. The AI system will send clips and recommend how to act, but it relies on the operator.
Allison said he suspects that surveillance staff will remain the same but their job will change and become more analytical and investigative. There may be position increases.
“We don’t know how many payer errors have been happening over the last 50 years and how much cheating is going on,” Allison said. “My gut feel is table games will be able to reduce their supervisor levels by 100%, and that’s going to be a huge savings. They are there to rate players, and they won’t need that anymore. I believe the ROI of this will be the reduction in levels of staff on the gaming floor.”
