Someone has to do it

Warning: The contents of this story are not for the delicate of constitution.

The flickering screen shows a crude picture of a naked woman. The twenty-something woman who’s viewing it looks slightly bored, as though she’s seen the same image hundreds of times before.

The guy sitting next to her is perusing a gambling Web site, while a colleague sitting opposite her is clicking from a shopping site to a political advocacy page. The room they’re in could be mistaken for an Internet cafe‚ in any big city with young people sitting side-by-side at tables separated by desk-mounted screens. Instead, these folks are at work categorizing Web sites for Internet filtering firm Websense Inc.

Websense is nestled among the other high-tech and biotech firms that surround the vast, hilltop campus of the University of California, San Diego. Each day the company’s 24 analysts sit in a room at the back of the Websense building trawling through Web sites to update a database of more than five million sites that are classified into 80-plus categories. Customers receive the updated list every night and use it to either block employee-access to specific sites or to monitor whether those URLs show up in their logs.

“My friends think it’s funny I get paid to look at porn sites,” says Alex Hirsch, who has been working as an analyst for six months. “I like my job because I get to use my languages — German, French and Italian,” she says.

Most of the analysts are overseas students at the university. They work a 20-hour week at Websense, which keeps its facilities open 24-7 so that the students can work a flexible schedule. Together, the analysts are fluent in 24 different languages, which enables them to post — classify sites according to language and category — in 52 languages because they can classify a Web site that’s written in an unfamiliar language by just looking at the content. The company also will hire contractors if its employees come across a load of sites in a language that no one understands.

But Patrick Swisher, director of database operations, says having a workforce with different cultural backgrounds is important so that proper judgment is used when categorizing certain sites. For example, according to Websense, legitimate business firms in Korea routinely use sex to lure visitors to their sites, which means they shouldn’t be categorized under pornography.

Some porn sites are disturbing compared with some of the other stuff the analysts have to view and categorize. Olga Bernasorskaya, a marketing and public relations student at the university who speaks several Eastern European languages, has worked at Websense for the past two years. She spent two months on a project researching child pornography sites.

“I said to myself this was just work, but yes, it was disturbing,” she says. “If I saw a naked child, I would classify it straight away. I looked at a kid on one site and I knew it was a Russian kid — it was very sad.” She says it felt good to work on the project because she helped to identify 500 pedophile sites that Websense customers can block.

Swisher says the analysts take their work in stride and that only two employees have said they felt uncomfortable. “Once you’ve seen 20,000 pairs of breasts, another pair is ‘So what,’ ” he says, adding that the sites to be classified are spread evenly among the analysts and that it is not possible to reduce the number of offensive material an analyst might have to view.

“Yes, there are some goofy sites out there. In our pre-screening phone calls (with prospective employees) we tell them the job will be viewing anything on the ‘Net and ask whether they would be comfortable with that,” he says. Employees are asked to sign a release form cautioning them that they will be viewing some offensive sites, and visitors and staff are greeted with a large warning sign on the door to the room where the analysts work.

Each day, the analysts categorize between 3,000 and 5,000 URLs that were either sent by customers or found using Websense’s automated search facilities. The first sites that get categorized are those likely to fall within Websense’s “obscenities” list. The list, which was previously called the Sinful Six until Websense acknowledged that one culture’s sin might not be regarded as sinful in another, includes adult content, weapons, race/hatred, illegal drugs, violence and tasteless.

There are also two other types of URLs: Premium groups (instant messaging, online day trading, paid to surf, streaming media, spyware) and sites most frequented. The analysts are assigned sites to classify according to the languages they speak and the priority order of obscenities, premium groups and most visited. The analysts classify the sites under one of more than 80 categories, including advocacy groups, education, and news/media.

It’s very hard for some visitors to the database room not to feel uncomfortable seeing these young people view some very graphic material. One employee says he focuses his gaze straight ahead when he walks through the area. But Websense tries to make the room feel comforting; on one side of the wall an artist has painted a mural of famous world sights, including London’s Big Ben, the Golden Gate Bridge and Sydney Opera House.

The analysts take great pride in their work, which requires patience and focus. Analyst Hirsch says she enjoys unearthing sites that try to hide the true nature of their content. “You see a beach scene so it could be a travel site, but you go further and you see it’s clearly a porn site. They hide it because they don’t want to be noticed by companies like us,” she says. But Websense wants people to know that it is merely the messenger when it comes to its customers’ Web access rules.

Tim Lan, a database services representative, has to sift through piles of, at times, abusive e-mails from people who tried to access a juicy site but instead were met with an all-white access-denied box, complete with Websense’s logo. “We tell them it’s their company’s policy that they can’t access that site, and that they should see their network admins,” he says.

Related Download
CanadianCIO Census 2016 Mapping Out the Innovation Agenda Sponsor: Cogeco Peer 1
CanadianCIO Census 2016 Mapping Out the Innovation Agenda
The CanadianCIO 2016 census will help you answer those questions and more. Based on detailed survey results from more than 100 senior technology leaders, the new report offers insights on issues ranging from stature and spend to challenges and the opportunities ahead.
Register Now