Employee Monitoring

By Larry Barrett Print this article Print

Non-obvious relationship awareness" (NORA) software probes databases in any language, searching for obscure matches between relevant information. Anonymized data (ANNA) software uses the same technology to investigate data that has been encrypted.

Employee Monitoring

From readings of employees' hands, eyes or other features to trackable tags on security guards to deep searches of databases, Las Vegas casinos use both their own intelligence and analytical software to check the backgrounds and activities of current and future employees.

"People change and their circumstances change all the time," says Shepherd, the security director at the Venetian. "It's not enough to do a battery of background or credit checks on an employee before you hire them. You have to know what's going on with them right now and six months from now."

Casinos use databases that include the names of suspected terrorists, felons, money launderers and child molesters provided by the FBI, private investigators and local law- enforcement agencies to screen employees throughout their tenure. They also apply NORA software, for unexpected clues.

It's not unusual, for example, for a casino to use NORA to discover a blackjack dealer who recently filed for bankruptcy protection or a cashier who was arrested for possessing narcotics an hour after finishing his shift by collecting arrest information from the Las Vegas Metro Police Department. If the end-of-the-shift count at the cashier cage shows $200 is missing or if players appear to win at a disproportionate rate during a particular dealer's shift, the investigation might begin with those wrestling with off-duty demons.

These details, including any changes in residence or marital status, are continually updated on a casino employee's security profile.

Shepherd says the Venetian also plans to eventually use tags that can be read by radio waves to track security officers and equipment throughout the casino, rather than rely solely on video surveillance and radio communication.

At the north end of the Las Vegas strip stands the Stratosphere Casino Hotel & Tower. It features a 1,149-foot observation tower the hotel claims is "the tallest building west of the Mississippi River."

Perhaps that was one attraction for Mohamed Atta and four other terrorists when they stayed in a hotel across the street a few months before they attacked New York's tallest buildings, the towers of the World Trade Center.

But if that was enough of an attraction that they actually entered the Stratosphere, Atta and his cohorts would certainly have been caught on camera. If the hotel had been told to watch out for individuals that looked like any of the five, they likely would have been noticed as they wandered onto the casino floor. That's because the Stratosphere has been using facial-recognition software to monitor its casino operations for more than 5 years.

Facial-recognition software mathematically measures features such as the distance between eyes or the structure of a cheekbone to compare a face to stored images.

The Stratosphere figures its $7,000 system paid for itself in its first month of operation. But that's hard to prove, says Roger Williams, a former member of a team of "card counters"—gamblers who use statistical methods to beat the house at blackjack.

"The truth is that there is no way to audit these systems to see if they are really performing," Williams says. "And both the casinos and the companies selling them have a large incentive to lie about their effectiveness."

Card counting isn't illegal, but casinos have the right to keep gamblers employing the technique from playing or to eject them from the premises. So Williams, the members of his team and their leader, known as "Mr. X.," keep a low profile when visiting Vegas.

The key, according to Williams, "is to not be noticed. Once you are noticed, you are toast; it doesn't matter whether they have a biometric ID [of you] or not."

To that end, Williams says "X" never stays at one casino for too long, and changes his appearance regularly to avoid being recognized by casino employees. "When he does come back to a place after a few weeks or months, he has different hair, clothes, et cetera. The key is to be just another businessman on vacation. Vacations don't last long, so neither does X's play in a particular guise," Williams says.

Facial-recognition systems are designed to see through disguises such as the ones X employs. Indeed, the Stratosphere believes it has saved $492,000 in losses each year that its system for recognizing faces has been in operation, says Derk Boss, its surveillance director. The casino also saves $44,000 each year in labor costs that otherwise would have gone to checking out the identities of players. All for a one-time cost of $7,000 and a few hundred dollars a year to subscribe to database and software updates and an Internet service that connects Stratosphere to the surveillance rooms at over 170 other hotels.

The Stratosphere uses a collection of software called Visual Casino from Biometrica Systems, the Las Vegas-based subsidiary of Viisage (see Dossier, p. 54). The software components compare images captured from casino surveillance video against a set of "mug book" databases and offer a selection of possible matches. "The software doesn't give an 'exact match,'" says Biometrica's Pepin. "It brings up the most likely matches, ranked from highest to lowest, in sets of nine."

Visual Casino uses two sources of data for facial recognition: a CD-ROM database from Biometrica called Casino Visual Identification; and an in-house database built from the casino's own intelligence gathering, using the same tools and formats as the visual database. Each entry in these databases includes known associates, a collection of images of the individual (especially if he or she is known to use disguises) and a description of their modus operandi.

But the matches that the Stratosphere gets from its software are only as good as the people using it. According to Greg Shanton, the chief technology officer and security director for system integrator American Management Systems' security group, computers alone aren't enough. "You have to have trained operators making the decisions," he says.

How This Applies to Domestic Security: Facial recognition currently is being used only at a handful of airports, only at security checkpoints and generally by lightly trained Transportation Security agents. If casinos' practices were followed, images would be taken at check-in, while would-be passengers obtain boarding passes and check luggage. This would give surveillance agents more time to match faces with known identities as well as to check those identities against the driver's license and credit card presented by each passenger—and crosscheck with watch lists and other databases.

This article was originally published on 2004-04-04
Senior Writer
Larry, of San Carlos, Calif., was a senior writer and editor at CNet, writing analysis, breaking news and opinion stories. He was technology reporter at the San Jose Business Journal from 1996-1997. He graduated with a B.A. from San Jose State University where he was also executive editor of the daily student newspaper.
eWeek eWeek

Have the latest technology news and resources emailed to you everyday.