'Brutal gut-punch': Report details new national security threat posed by China
Artificial Intelligence that's being built by private labs to store sensitive U.S. national security information is at risk of being hacked by the Chinese government, according to reporting from TIME.
That's because components used to build the technology "are mostly or exclusively built in China," wrote correspondent Billy Perrigo.
Perrigo reviewed a government-sponsored report titled America's Superintelligence Project, that was released to the public Tuesday as a redacted version.
The report's authors with consulting firm Gladstone AI, surmised that "today’s top AI datacenters are vulnerable to both asymmetrical sabotage—where relatively cheap attacks could disable them for months—and exfiltration attacks, in which closely guarded AI models could be stolen or surveilled," Perrigo wrote.
ALSO READ: 'Dictatorship, not a town hall': Families 'distraught' as MTG disruptors tased and jailed
The authors told TIME that "even the most advanced datacenters currently under construction—including OpenAI’s Stargate project—are likely vulnerable to the same attacks."
“You could end up with dozens of datacenter sites that are essentially stranded assets that can’t be retrofitted for the level of security that’s required,” said Edouard Harris, one of the authors of the report. “That’s just a brutal gut-punch.”
The authors speculated about a potential attack "that could be carried out for as little as $20,000, and if successful could knock a $2 billion datacenter offline from between six months to a year."
When U.S. tech labs order parts to fix the AI hampered by an attack, "China...is likely to delay shipment," the report said.
Perrigo wrote that "the report addresses calls from some in Silicon Valley and Washington to begin a 'Manhattan Project' for AI, aimed at developing what insiders call superintelligence: an AI technology so powerful that it could be used to gain a decisive strategic advantage over China."
But, the authors warned, “if we’ve actually trained a real superintelligence that has goals different from our own, it probably won’t be containable in the long run.”