The mass data collection combined with better data analysis tools translates into an "arms race" where intelligence officials try to find new connections with the data they collect, said Ashkan Soltani, a technology and privacy consultant. New data analysis tools lead intelligence officials to believe they can find more links to terrorism if they just have "enough data," but that belief is "too much sci fi," he said.
"This is the difficult part, if you're saying that if we have enough data we'll be able to predict the future," the ACLU's German said.
Many U.S. intelligence officials are suspect of tech vendor claims about predictive analysis, Edgar countered. However, link analysis — the tracking of suspects through communications with other known criminals or terrorists — is a "very powerful tool," he said. It may be possible to use sophisticated cryptographic techniques to do that kind of analysis without the bulk collection of phone records, he said.
"The [internal] compliance regime is not the best answer for privacy," Edgar said. "The best answer is not to take the data in the first place, then you don't have to worry about compliance."
The ACLU has concerns about link analysis, because it creates a massive list of suspicious people that overwhelms investigators, German said. "What link analysis creates is suspicion upon the people that suspicious people are linked to," he said. "That growing cloud of suspicion can never been cleared."
Sign up for CIO Asia eNewsletters.