'Digital Silhouette' Capability Raises Privacy ConcernsAdvertising and e-commerce tracking company Predictive Networks Inc., Boston, opened for business this week, promising prospective clients that it can track every site a Web surfer visits and build a profile, or "digital silhouette," of online behavior.
Of course, privacy advocates have long criticized such practices. But dissatisfaction with highly intelligent profiling technology like Predictive's is beginning to annoy academic leaders who fear that automated profiling of people by corporations is becoming a standard. And they apparently see little difference between telephone keypads and computer keyboards.
Dave Farber, a privacy advocate and professor of telecommunications at the University of Pennsylvania, expressed his dissatisfaction with the new technology by saying last week in a published report, "It takes a court order and tight supervision to listen in over a telephone link."
But according to Gus Bickford, spokesman for Predictive Networks, the company is hardly listening to people's conversations.
"From day one, we have focused as hard as possible on privacy," he said. "We do not sell data to advertisers. What we do is provide advertisers targets to sell to via our servers. Our information is proprietary, but it's really a whole series of different vectors -- no personally identifiable information."
Indeed, Internet advertising networks of all kinds argue they are not tracking people's actual "conversations" or reading their e-mails and chat messages. They usually are amassing aggregate data such as browser cookies and Internet service provider traffic information as a means of analyzing sales prospects for clients. Those assurances, however, still are getting lost on most consumers.
Clayton said, "Consumers worldwide are demanding that businesses use personal information responsibly." He said his company recently formed a strategic alliance with IBM, Armonk, NY, and P&A Advisors LLC, Boston, to provide those privacy solutions to clients.
But Predictive Networks characterizes its platform as one that uses an advanced, artificial intelligence technology engine that is highly scalable and is part of a fault-tolerant network. The company claims the engine "learns from online behavior to deliver intelligent, well-timed, desired content." Based on this learned information, Predictive Networks said clients can build or provide more uniquely targeted Web content, advertisements, tailored news and sports updates, network announcements and so on.
Predictive Networks claims it only analyzes click-stream data to develop an in-depth understanding of individual Internet subscribers' interests and characteristics, then sends those subscribers personalized information tailored to their needs and predicted receptivity.
Even the company's Web site promises consumers, "It's all done anonymously, without any effort on the part of the subscriber. No surveys. No cookies. No personally identifiable information."