What does Tor's allegation against the FBI mean for computer research?

Representatives of the Web anonymity service The Onion Router, better known as Tor, this week accused the FBI of paying researchers at Carnegie Mellon University to unmask Tor users. The allegation, which an FBI spokesperson denied, has alarmed some cryptographers and raises questions about the collaborative relationship between law enforcement and security researchers.

Matthew Green, a computer scientist who teaches cryptography at Johns Hopkins University, told FCW that the FBI's alleged hiring of CMU researchers crossed the line of generally accepted collaboration between academia and government on computer security research.

"Most people in our field would do one thing outside of the university…but within their university research they're pretty careful about what they do, with very few exceptions," Green said.

Court documents posted by Motherboard state that a member of Silk Road 2.0, the now-dismantled illicit online bazaar, and a man charged with possessing child pornography were identified with the help of "university-based research institute that operated its own computers on the anonymous network used by Silk Road 2.0."

Tor Project representatives say that research institute is CMU's Software Engineering Institute, and that it took $1 million from the FBI to target Tor users. Among the evidence supporting the allegation is the similarity between the attack on Tor users and a presentation that two researchers at the institute's CERT division submitted to Black Hat and later withdrew.

"We think it's unlikely [the FBI] could have gotten a valid warrant for CMU's attack as conducted, since it was not narrowly tailored to target criminals or criminal activity, but instead appears to have indiscriminately targeted many users at once," said a blog post from the Tor Project.

A FBI spokesperson denied the charge, telling FCW: "That allegation that we paid Carnegie Mellon $1 million to hack into Tor is inaccurate." The spokesperson would not comment when asked if the FBI enlisted Carnegie Mellon's researchers directly to conduct the attack on Tor.

Tor Project representatives could not be reached to comment. A CMU spokesperson declined to comment.

The attack was on Tor's hidden services and not the Tor browser itself, which have two different security levels.

"It's important to distinguish between hidden services and the Tor browser -- hidden services are just the transport layer for ordinary Web servers," a former Tor employee told FCW. "So if you're running unpatched software to deliver JavaScript or Flash, you're going to have a bad time."

There is a genuine risk that the attack exposed an untold number of innocent Tor users, according to computer security experts.

The withdrawn Black Hat presentation "suggested that the researchers basically tried to de-anonymize all users of all hidden services," Nicholas Weaver, senior researcher at the International Computer Science Institute in Berkeley, Calif., told FCW. The "flow marking" for tracing network users "was also potentially visible to others," he added.

Green said CMU researchers "didn't take any special precautions to make sure that that information was only visible to them and not other people," such as encrypting that information.

A sophisticated set of computer specialists, such as those backed by a nation-state, could have noticed the allegedly CMU-backed attack going on and "piggy-backed on this attack and thousands of people who are using the system could potentially have been de-anonymized," Green added.  Human rights activists and journalists rely on Tor to shield their digital footprints from authoritarian governments.

Though the FBI has invested considerable resources in building out its own cybersecurity personnel, Green said the complexity of the field and its pace of change mean the Bureau likely will need to continue to partner with academia. To what extent this episode sullies the well of collaboration remains to be seen.