G2E: New generation of facial recognition enhances security, raises questions

October 21, 2019 9:37 AM
  • Mark Gruetze, CDC Gaming Reports
October 21, 2019 9:37 AM
  • Mark Gruetze, CDC Gaming Reports

Casinos finally can get facial recognition technology that works, but operators must be prepared to handle all the information it will provide, a Global Gaming Expo panel said.

Story continues below

In the past couple of years, the technology has advanced dramatically, providing security experts with the ability to identify violent criminals or blacklisted patrons before they enter a venue or offering marketing executives the ability to obtain even more revealing data about players.

But the technology also could end the “didn’t know/can’t tell” defense for not identifying suspected “chip walkers” and others involved in potentially questionable transactions.

“The reward side is probably the sexier topic for a lot of folks in gaming,” said Nasr Sattar, vice president of NRT Technology Corp.’s Innovation Group. “Compliance is also a very risky area. No one wants to get a million dollar fine for not knowing who that person is and not reporting it.”

Sattar spoke Thursday at a G2E panel discussion on “Customer Identification Using Facial Recognition Technology: The Future is Now.” Also on the panel were Jessica Medeiros Garrison, president of MDM27 Holdings, whose company Clearview offers facial recognition technology to law enforcement agencies, and David Logue, vice president of security, surveillance and nightclub compliance for the Cosmopolitan of Las Vegas. Alec Massey, director of PwC Connected Solutions, moderated.

Facial recognition works by mapping facial features from a photograph or video, and then comparing the information with a database of known faces to find a match.

Massey said the introduction of facial recognition programs three to five years ago worked in labs but failed in the real world.

Now the technology is evolving quickly, with casinos, hotels, and other retail establishments converting to the technology, Massey said. Last month, the U.S. Traffic Safety Administration completed a 30-day test at McCarran International Airport in Las Vegas using facial recognition for verify boarding passes; results are to be released next year.

“We moved from a place where technology was the barrier to one where regulation and privacy concerns are the primary barriers to what’s holding facial recognition back,” Massey said.

Medeiros Garrison said any facial recognition program more than three or four years old is “definitely a super-crappy, old algorithm.”

She said current versions rely on recent improvements in artificial intelligence and machine learning that can account for differences in viewing angles and lighting. She said online security verification tests that require a user to identify which of several pictures contain a stop sign, for example, are a way of training an artificial intelligence algorithm.

Logue said the most immediate use for facial recognition is in keeping venues safe because it can scan people in a crowd and quickly match hits on people with criminal or violent backgrounds.

“If you can identify them right then and stop them, think how much safer your employees and patrons and everybody is,” he said. In Nevada, casinos and other venues have the right to refuse entry in such a case.

He said the system would require databases with hundreds of thousands of images, and a need to hire additional people to manage the databases and confront those identified as risks.

Sattar said facial recognition can be vital to a casino’s handling of its own risks and rewards.

The risk side includes the reporting and compliance requirements for all types of casinos transactions.

He said one use is with someone suspected of being a chip walker, a patron who fails to cash a large amount of chips, presumably to be used for illegal payments outside the casino.

“You’ve associated that patron with a face. That goes with your compliance tools,” he said.

“In the old days, if you didn’t know, you didn’t have to report it. That’s rapidly changed.”

On the reward side, Sattar said facial recognition could eliminate gambling’s “last black hole of information:” rating table-game players.

Instead of ratings based on brief observations by harried pit supervisors using pencil and paper, facial recognition not only would recognize the player but use technology to track bet size, side bets placed, and other factors.

“There are all sorts of implications that come out of that,” he said. “You get player utilization, you understand your side bets a lot better, you understand your base bets a lot better. And finally, you understand who your players are and how you want to market to them.”