Google’s Security and Resilience Services (GSRS) team will use the data to help identify people “who may pose a security risk to Google’s people products, or locations,” the document says.

“There are protocols in place for identifying, reporting, and potentially removing known unauthorized persons to maintain safety and security of our people and spaces,” it says.

At the Kirkland testing site, people entering the building will not be able to opt out of the facial screening. However, the document says the data is “strictly for immediate use and not stored,” and that employees can opt out of having their ID images stored by filling out a form. Google told CNBC that while ID badge photos were part of the test, they won’t be used going forward.

“For many years our security team has been testing and implementing new systems and protections to help keep our people and spaces as safe as possible,” a Google spokesperson said in an email.

Google has experienced at least one notable violent incident in the past. In 2018, a woman opened fire at YouTube’s office in San Bruno, California, injuring three people. The shooter allegedly targeted YouTube because she “hated” the company for blocking her videos.

The Kirkland test lands at a sensitive moment for Google, which is at the center of the artificial intelligence boom and is rapidly adding AI across its portfolio of products and services. Facial recognition technology is particularly controversial because of the privacy concerns around surveillance.

In 2021, Google proposed new security changes, including fences around parts of its headquarters in Mountain View, California, especially as its construction plans included public and retail spaces. More recently, company executives have cited security reasons for cutting off access to employees after a series of layoffs and protests over the past year.

In early 2023, the company announced plans to eliminate about 12,000 jobs, or 6% of its workforce, in response to a downturn in the online ad market and a broader economic slowdown. Google has laid off more employees recently, moving some engineering roles to India and Mexico.

In a high-profile incident in April, Google terminated more than 50 employees after a series of protests over labor conditions at the company and against Project Nimbus, Google’s cloud and AI contract with the Israeli government and military. Employees staged a sit-in protest at offices in New York and Sunnyvale offices.

Chris Rackow, Google’s vice president of global security, told staffers at an all-hands meeting last month that “extensive use of all of our video camera footage” helped to identify employees that the company said were disruptive during the protests and who made their colleagues feel threatened and unsafe, according to audio of the meeting obtained by CNBC.

Facial recognition technology became a big topic for lawmakers in 2020, following pressure from civil rights advocates and national protests sparked by the murder of George Floyd. Amazon, Microsoft and IBM imposed restrictions on the sale of their technology to police.

The following year, Amazon was questioned by U.S. senators about its use of employee surveillance after the company deployed AI-equipped cameras in delivery vans. In April, warehouse workers sued Amazon alleging the company illegally collected biometric data that included face scans. And late last year, the Federal Trade Commission proposed barring Rite Aid from using facial recognition software in its drugstores for five years to settle allegations it improperly used the technology to identify shoplifters.

Security is a costly endeavor for Google not just on campuses but all the way up to the top ranks of the company. In 2023, CEO Sundar Pichai’s personal security cost the company $6.8 million, up from $5.9 million a year earlier, according to regulatory filings.

WATCH: Google, Microsoft announce layoffs