Politics
Anthropic Executive Tells Tennessee AI Council to Brace for 'Tumultuous Cyber Era'
Anthropic's head of geopolitics warned the Tennessee AI Advisory Council that AI-powered cyberattacks will surge over the next six to 24 months, urging state and local governments to harden critical infrastructure and deploy open-source defense tools.
NASHVILLE â A top executive at AI developer Anthropic told Tennessee's AI Advisory Council on Monday that state and local governments should prepare for a sharp rise in cyberattacks powered by artificial intelligence, warning that the window to shore up defenses is closing.
"We really feel we are entering into a pretty tumultuous cyber era over the next six to 24 months," Thompson Paine, Anthropic's head of geopolitics, told the council during a Nov. 16 presentation. "We want to make sure that we are protecting American critical infrastructure as best possible."
The council, composed of state business and government leaders, heard Paine outline a "coming wave of cyberattacks" that he said could be blunted through cyber defense programs, focused protection of power, water and transportation systems, and broad deployment of open-source software capable of running free, recurring security scans.
Paine's central concern is that AI has lowered the barrier to entry for would-be attackers. "AI is giving not just traditional attackers advanced capabilities; it's giving anyone advanced capabilities," he said. "I don't need to know anything about writing code to build software. That means I also don't need to know anything about cyberattacks to execute a cyberattack now. I think that's where we're very concerned."
The challenge, he said, lies in finding vulnerabilities in existing systems and fixing them quickly.
Paine also addressed the intensifying U.S.-China competition over AI. He pointed to Anthropic's own reporting last year that Chinese hackers used its Claude chatbot to target roughly 30 companies and government agencies â an operation widely regarded as the first cyber espionage campaign carried out with AI. He estimated the U.S. holds about a six-month lead over China in AI, noting that while China has strong talent and cheap, abundant energy, it lags in domestic compute capacity.
"They are several years behind us on the ability not just to produce chips at the same quality of [AI] chip companies in the U.S. like Nvidia, but also at the same quantity," Paine said, adding that federal restrictions on China's access to U.S.-made chips would widen the gap further. Nvidia said Wednesday it would stop shipping certain chips to China.
Tennessee Attorney General Jonathan Skrmetti, a council member, questioned Paine about the trajectory of the China rivalry.
The presentation came during a consequential week for Anthropic. On Wednesday the company sold more than 956 million shares at $83 apiece, setting a record for an initial public offering and raising approximately $79.4 billion. Anthropic later said other stockholders sold more than 23 million shares, bringing total proceeds to more than $83.1 billion. The offering was oversubscribed, with both the share count and price exceeding the company's initial expectations of 825 million shares at $70 to $78 each.
The warning also lands amid a nationwide uptick in cyberattacks and communications outages, including incidents at Microsoft, United Airlines and Apple, at agencies such as the Federal Aviation Administration, and at institutions including Vanderbilt University, which has been grappling with a ransomware attack that knocked out cloud storage services.
For Tennessee, the message was pointed: the same AI tools that promise efficiency for government can be turned against it, and the state's utilities, water systems and transit networks are among the assets most exposed.