US military partners with AI to crack down on ‘disinformation’
US Special Operations Command (USSOCOM) last week partnered with Accrete AI to fight disinformation on social media, the company announced.
USSOCOM will use Argus, an AI program developed by Accrete, to predict and detect disinformation. Argus is an anomaly detection tool which analyzes online behaviors to find inconsistencies. The company says Argus can detect generative AI and other “synthetic media.”
“Synthetic media, including AI-generated viral narratives, deep fakes, and other harmful social media-based applications of AI, pose a serious threat to U.S national security and civil society,” said Accrete Founder and CEO Prashant Bhuyan in a statement. “Social media is widely recognized as an unregulated environment where adversaries routinely exploit reasoning vulnerabilities and manipulate behavior through the intentional spread of disinformation.”
Bhuyan added that Argus will help USSOCOM “predict social media narratives at an embryonic stage” and suppress them before they “evolve and gain traction.”
The USSOCOM is an additional contract to the one last year between Accrete and the Department of Defense (DOD). In November 2022 Accrete announced it had been awarded a five-year multimillion dollar contract by the DOD for Argus, which had already received strong approval from the Pentagon in 2020.
But while some generative AI may indeed pose risks to national security, the contract comes after a judge found the Biden administration to have perpetrated “the most massive attack against free speech in United States’ history” under the pretext of “disinformation” and “misinformation.” US District Judge Terry Doughty, who said it was “telling that each example or category of suppressed speech was conservative in nature,” prohibited the government from further colluding with social media companies to suppress content online.
The Biden administration has appealed the ruling, arguing the need to censor Americans on social media for “national security.”
In addition to Accrete, the Biden administration has already begun working with major corporations on AI development. Executives from Microsoft, Google, Meta, Amazon Web Services, OpenAI, Inflection AI and Anthropic met with White House officials in July to pledge their commitment to new dicta set by the federal government. While the commitment is voluntary for now, it is expected to be mandated through an executive order in the coming months.
Accrete also plans to launch a similar “disinformation detection tool” for businesses later this year.
