A new cyber defense system being developed by the NSA could automatically launch counter-strikes against attackers who target the US, whistleblower Edward Snowden has claimed.
The MonsterMind project, still under development at the spy agency, features algorithms which would automatically scan vast chunks of metadata with the aim of picking out malicious traffic.
With that intelligence the NSA system could then neutralize the threat and even theoretically launch a retaliatory strike autonomously, Snowden told Wired.
However, such a capability could end up targeting the innocent compromised computers being used by an attacker as a botnet to launch the initial threat, the whistleblower cautioned.
“These attacks can be spoofed,” Snowden told the site.
“You could have someone sitting in China, for example, making it appear that one of these attacks is originating in Russia. And then we end up shooting back at a Russian hospital. What happens next?”
The second issue is that for the system to work effectively, the NSA would have to gain access to all communications traffic coming into the US. Seizing private comms without a warrant and with no suspicion of wrongdoing would violate the Fourth Amendment, Snowden added.
Sean Sullivan, security consultant at F-Secure, agreed that the MonsterMind may end up counter-attacking botnets comprised of compromised computers belonging to US citizens or allies of the States.
“Counterattack options are only useful if the adversary has something to lose. Take North Korea as an example,” he continued.
“It might attempt to launch an attack from comprised resources. But even if it used its own servers to attack US infrastructure – what besides those servers is there to counterattack? North Korea isn’t wired – it basically has nothing to lose.”
Sullivan labelled it an “overly complicated defense strategy”.
“A fraction of the money used by ‘MonsterMind’ could be spend on bug hunting and eliminating vulnerabilities to achieve greater results,” he told Infosecurity.