We may be a little distracted from what Anthropic is saying about artificial intelligence and weapons, surveillance, and cyberwarfare by what a former Anthropic employee, Jacob Coxon, said on his way out the door last week.
“These will soon be superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources,” Coxon tweeted on X. He added: “The people building AI earnestly believe that it could kill us all by the end of the decade,” a claim that several Anthropic employees have said is accurate.
An Anthropic spokesman’s official response: “We have always been transparent that AI will bring both enormous benefits and unprecedented risks. This work is also why we believe the world would benefit from the industry adopting a lawful, verifiable way to work together to pace how we release powerful models.”
You’d think that the Anthropic people have almost nothing to do with what is being done within one of the major AI companies, Anthropic, that they are talking about. If more attention needs to be paid to safety, how about paying more attention to safety? Maybe the response would be that Anthropic is doing all it can and cannot be expected to slow down unilaterally.
In conclusion, mankind will or will not end by 2030.
Lesser AI destruction
Assuming that mankind will not end, or at least that we should try to have as good a world as possible during the four or five years that are remaining to us, let us now consider the problem of AI’s aiding and abetting of authoritarian and totalitarian “weapons, spying and cyber operations” (Reuters, September 11, 2026).
Anthropic said a China-based actor used Claude to develop an electronic-warfare and air-defense suppression software suite that ranked targets and modeled radar jamming. During the project, the actor changed a simulation to include 12 targets in Taiwan, including early-warning radar, Patriot and Tien Kung missile batteries, air bases and a command bunker. Anthropic said account information linked the actor to Chinese research institutions including the People’s Liberation Army Academy of Military Sciences. Anthropic banned accounts linked to the actor….
Anthropic said another China-based actor used Claude to help develop specifications and fire-control software for an anti-torpedo system intended for the Chinese navy. The actor used Claude to produce a technical proposal of more than 200 pages, compare the system with U.S. Navy technology and simulate hostile technical reviews to improve the proposal. “We assess the actor was associated with a Chinese defense industry manufacturer aiming to produce a weapons specification and acquisition proposal for the People’s Liberation Army Navy,” Anthropic said.
Various China-based actors used Claude to research microwave weapons, launch cyberattacks on “about 50 organizations, including foreign government networks,” track Uyghurs in Syria, and “compile intelligence on targets including Catholic cardinals, Taiwanese Christian leaders, Tibetan Buddhists, dissidents and activists.”
Other countries cited for their maleficent use of Claude include Russia, Iran, Yemen, and Mali.
Anthropic’s report on the last eight months of major misuse of Claude says that the firm’s Threat Intelligence Team both identified and disrupted these major operations. Judging by the examples, though, the disrupted operations were often disrupted only after they were well underway. The 200-page technical proposal to improve a PLA Navy anti-torpedo system went through “many drafts” before Anthropic detected this work and banned the account.
“The cases we share here aren’t typical misuse, but rather examples of the most notable and novel threat activity we’ve identified to date.”
And: “As models become increasingly capable, their risks will increase, unless AI developers and society’s defenders act to make them safer.” We seem to have circled back.
Also see:
Anthropic: “Detecting and countering misuse of AI: September 2026”
”Sophisticated and persistent threat actors continuously test our safeguards and try to circumvent the technical measures we use to detect and prevent misuse. We’ll continue to evolve [improve?] our safeguards and coordinate with our partners to improve our ability to detect, disrupt, and prevent future misuse.
“We hope that the findings in this report will help other developers recognize similar patterns on their own platforms, give governments and civil society a clearer view of how emerging threats take shape, and strengthen collective defenses.”
Dario Amodei: We Must Pace the Frontier (September 2026)
“Not building the technology deprives humanity of benefits or simply places AI in the hands of authoritarian powers, while building it too fast is reckless.”
