AOL Releases Web-Search Data Of 650,000 Users By RIVA RICHMOND Wall Street Journal
August 8, 2006; Page A2 http://online.wsj.com/article/SB115498329554829021.html?mod=Technology AOL Inc., in a move that raised online-privacy concerns, said it mistakenly released data about the Web-search habits of more than 650,000 AOL members. The AOL users weren't personally identified in the data and instead were tracked by anonymous user-ID numbers. However, numbers would still allow everyone from law enforcement to identity thieves to analyze an individual's searches -- which could involve names, addresses and other subjects that could provide hints to the individual's identities. Andrew Weinstein, a spokesman for the Time Warner Inc. unit, acknowledged that "search queries themselves can sometimes include such information" and called the release "a screw-up." AOL researchers posted the data, which detailed more than 20 million queries made by the users between March 1 and May 31, without authorization, to a new AOL research Web site about 10 days ago. The document was for use by other search-technology researchers, but was noticed by bloggers in the search-marketing field late Sunday. Their discovery set off a flurry of blog postings and apparently led to hundreds of downloads of the data. AOL said it immediately pulled the data off its site Sunday when it realized what had happened. "We're angry and upset about it," AOL spokesman Andrew Weinstein said in an emailed statement. "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." The company is conducting an internal investigation before determining whether it will take disciplinary action against employees. The incident recalled a controversy sparked earlier this year by Google Inc. when it refused to turn over to the Justice Department a sample of anonymous search queries conducted by its users, citing competitive concerns and consumers' privacy fears. The government sought search data from Google and other major search engines for use in its defense of the Child Online Protection Act. A federal judge ruled in March that Google didn't have to provide the data. While AOL did cooperate with the Justice Department, the company said yesterday that it didn't provide user IDs akin to those used by the AOL researchers. Indeed, AOL's recent data release is more consequential because the data set was very large, user anonymity was less assured and the disclosure was made to the world, not one government agency. Indeed, the data continue to circulate actively on the Web. "I'm gratified that they understand this was a mistake," said Kurt Opsahl, an attorney at the Electronic Frontier Foundation, an online-privacy advocacy group. "But once the information is out there it stays out there. The horse is out of the barn." Search data can reveal very personal information, including political and religious beliefs, medical conditions and financial information, he said. "It's a serious disregard for user privacy, particularly considering the uproar over the [Justice Department's] demand for this kind of information from Google earlier this year." The flub highlights the danger of allowing search companies to retain personal data and the need for clear online-privacy protections, he said. The Electronic Frontier Foundation is still evaluating the matter, but Mr. Opsahl said the disclosure "appears to violate their privacy policy, and it may violate federal law." "This act, I view it as a black eye for the entire industry," said Alexander Tuzhilin, a professor of information systems at New York University's Stern School of Business. "It makes all of us sort of concerned because there's so much personal information available online these days and it can happen to anybody," he said. Meanwhile, search marketers have been gleefully parsing the data for popular search terms and information about when users click on links on search-results pages. "After taking a quick peek at the data, its an absolute gold mine," one marketer wrote on his blog. He listed a string of "scary" searches by one user on terms including "how to kill your wife," "dead people" and "car crash photo." According to Web-measurement firm comScore Media Metrix, AOL's search network, which is powered by Google technology, served 42.7 million visitors in May, suggesting the disclosed data covered roughly 1.5% of its May search users. ================================ George Antunes, Political Science Dept University of Houston; Houston, TX 77204 Voice: 713-743-3923 Fax: 713-743-3927 antunes at uh dot edu Reply with a "Thank you" if you liked this post. _____________________________ MEDIANEWS mailing list [email protected] To unsubscribe send an email to: [EMAIL PROTECTED]
