Google Trades Blame with HLS

Google mistakenly blamed a little known Harvard organization this weekend for causing an error that briefly caused its search engine to warn users that nearly all Web sites contained potentially harmful malware, or software that could infect or damage a user’s computer without consent.

From 9:30 a.m. to 10:25 a.m. on Saturday, the search engine warned that clicking on most search results might “harm your computer.” The problem was fixed within an hour.

Initially, Google issued a statement on its official blog implying that a Harvard-affiliated internet watchdog group, StopBadware, was responsible for the malfunction.

StopBadware is a research venture of the Berkman Center for Internet and Society at Harvard Law School and is partly funded by Google. The organization works with the search engine to determine the criteria for identifying potentially harmful Web sites.

Google later backpedaled on its statement, conceding that the glitch had been caused by a simple human error. A forward slash was mistakenly added to a file that contained a list of potentially dangerous Web sites. As a result of this error, nearly every Web page was added to that list.

Max H. Weinstein, the manager of StopBadware, scrambled to correct the record, writing on his blog that “no data that we generate is supposed to affect the warnings in Google’s search listings.”

“Subsequent to that, after some e-mails and frantic phone calls, Google updated their statement to properly characterize their relationship with StopBadware and to correct that the error was on their end,” Weinstein said in a phone interview yesterday.

The warning typically informs Internet users that their computer might be infected with harmful software if they visit a suspect Web page, and it directs them to visit StopBadware’s Web site to learn more about how to protect their computers from malware.

As internet users around the world encountered the error page, traffic flooded StopBadware’s Web site. According to Weinstein, the Web site received at least three times its average daily hits. With such an increase in congestion, the site ended up crashing.

“I think it’s good that we raised awareness but I think it’s unfortunate that it had to be raised in this way,” said Weinstein.

By initially blaming a small, unknown organization for the error, it seemed like Google had hoped to transfer blame, but officials at StopBadware rejected that notion and said that Google’s quick response to the error was impressive given the complex nature of the search engine.

“In the heat of a very confusing situation somebody made a mistake and then very quickly retracted that statement,” said Law School Professor and co-founder of StopBadware John G. Palfrey ’94. “It’s the completely wrong reading that Google was looking for a scapegoat.”

Webmasters whose sites have been automatically blacklisted by Google can appeal to StopBadware to be removed from the list of flagged sites.

—Staff writer Elias J. Groll at egroll@fas.harvard.edu.