AOL apologizes for release of user search data

Search log information originally intended for use on new research site; company calls data posting a mistake.

AOL apologized on Monday for releasing search log data on subscribers that had been intended for use with the company's newly launched research site.

The randomly selected data, which focused on 658,000 subscribers and posted 10 days ago, was among the tools intended for use on the recently launched AOL Research site. But the Internet giant has since removed the search logs from public view.

"This was a screw-up, and we're angry and upset about it. It was an innocent enough attempt to reach out to the academic community with new research tools, but it was obviously not appropriately vetted, and if it had been, it would have been stopped in an instant," AOL, a unit of Time Warner, said in a statement. "Although there was no personally identifiable data linked to these accounts, we're absolutely not defending this. It was a mistake, and we apologize. We've launched an internal investigation into what happened, and we are taking steps to ensure that this type of thing never happens again."

Although AOL had used identification numbers rather than names or user IDs when listing the search logs, that did not quell concerns of privacy advocates, who said that anyone among the 658,000 could easily be identified based on the searches each individual conducted.

"We think it's a major privacy concern, and we're glad to see AOL is taking it seriously," said Ari Schwartz, deputy director of the Center for Democracy and Technology. "Companies that deal in search results have to understand that they carry very sensitive information, even if it doesn't have what we would traditionally consider to be personally identifiable information involved."

Schwartz and other privacy advocates noted that with bits of information, a "mosaic" could be created that could eventually lead a person to identify the individual in question.

related story
AOL offers glimpse into users' lives
Release of three-month search histories of some 650,000 users provides rare glimpse into their private lives.

"Sometimes what people are searching for may be an indicator of who they are and who they know," said Richard Smith, founder of Internet security and privacy consulting firm Boston Software Forensics.

In one search log, terms such as "how to tell your family you're a victim of incest," "casey middle school," "surgical help for depression," "can you adopt after a suicide attempt," "Fishman David Dr - 2.6 miles NE - 160 E 34th St, New York, 10016 - (212) 731-5345," "gynecology oncologists in new york city," and "how long will the swelling last after my tummy tuck" appeared in the set of data. (To see a more extensive account of search phrases surfaced in the AOL data, click here.)

Some researchers, however, contend the information serves a valuable purpose in helping to develop better information retrieval technology.

"Researchers at universities or small companies don't have access to this type of data. I think the (AOL) researchers were trying to do a good thing by making this available to the research community," said Steve Beitzel, who holds a doctoral degree in computer science from the Illinois Institute of Technology with a specialization in information retrieval. Beitzel, who is an affiliated researcher with the university's Information Retrieval Lab, once served as an intern at AOL, but was not involved with the release of the search log data.

In developing his doctoral thesis, Beitzel used another set of search data from AOL, unrelated to this recent issue, that focused on tracking trends in search query strings.

"It's a hot...research problem that people are trying to solve," he said.

Beitzel noted that the former Excite released a smaller data set of its users' search results in 1999 and 2001, and AltaVista engaged in a similar situation about five or six years ago.

Excite, as well as AltaVista, withheld the user's name and IP address and used an anonymous identifier.

"They released the data sets more than five years ago, and it hasn't hurt anyone," Beitzel said. "The bloggers say what AOL did was evil and a violation of privacy. But this may be an overreaction...a nine-digit number in a search box with no name attached is meaningless."

Kurt Opsahl, a staff attorney for the Electronic Frontier Foundation, pointed to other means to make the information available to the research community without making it open to the public.

"There are ways of conducting research into search technology, without making individuals' search terms public," Opsahl said. "Universities could abide by AOL's privacy laws and various laws for privacy...They could get consent from users before handing out the information to third parties."

While Beitzel agreed other methods could be enacted to aid researchers and the search community, he advised against issuing filters to screen out information such as names or Social Security numbers.

"If you alter the collection, then it is no longer representative," he said.

The release of the search logs runs counter to a court ruling in March, when a federal judge rejected efforts by the Department of Justice to gain access to Google users' search logs. The court, however, determined the Justice Department could have limited access to Google's index of Web sites.

Google was the only search engine to fight the Justice Department, with Yahoo, Microsoft's MSN and AOL turning over their users' search data.

"All search engines collect this kind of user data, and it's valuable to marketers, insurance companies, people involved in divorce and custody battles," said Rebecca Jeschke, a spokeswoman for the EFF. "If this information is available, there is a lot of temptation to release it."

Smith, meanwhile, noted the information AOL provided is similar to the type of search string information the Justice Department sought, under the Children's Online Protection Act.

The search log data, culled from March to May, represents approximately 1.5 percent of AOL's search network in May. The data applied only to U.S. searches by AOL subscribers using the company's client software.

A number of blogs are pointing to mirror sites to let people take a peek at the search logs of AOL users.

CNET's Declan McCullagh contributed to this report.

Featured Video
This content is rated TV-MA, and is for viewers 18 years or older. Are you of age?
Sorry, you are not old enough to view this content.

The WRT1900ACS is Linksys' new best Wi-Fi router to date

CNET editor Dong Ngo compares the new WRT1900ACS and the old WRT1900AC Wi-Fi routers from Linksys. Find out which one is better!

by Dong Ngo