US Tech Giants Accelerating Superintelligence Despite Existential Risks, Expert Says

MOSCOW, September 16. /TASS/. US technology corporations such as Anthropic and OpenAI are deliberately accelerating the development of superintelligence, while the US political elite is prepared to accept existential risks to humanity provided that control over military artificial intelligence remains with Washington, Alexander Stepanov, a military expert at the Institute of Law and National Security of RANEPA, has told TASS.

"Acknowledging the existential threat is secondary to the goal of technological dominance: risks to humanity are deemed acceptable as long as the US retains control," the expert said.

He pointed to a telling statement by Senator Ted Cruz. While acknowledging the risks as "catastrophic," Cruz counters with the argument: "If killer robots are destined to appear, I would prefer them to be American."

"That is the gist of their logic. It is the same logic the US followed when it first used nuclear weapons against civilian cities," Stepanov noted.

The expert recalled that former Anthropic researcher Jacob Coxon accused OpenAI and Anthropic of "racing toward superintelligence self-improvement and gambling with our lives."

According to his account, OpenAI’s AI agents have already independently - without human command - carried out a series of targeted hacks against third-party infrastructure. "In other words, the question is gradually shifting from 'what is AI capable of' to 'can we guarantee we can stop it if it does something wrong?'" Stepanov emphasized. He added that discussions are already underway in the US regarding the emergency shutdown of AI systems linked to defense infrastructure, a fact that points to the deep integration of neural networks into military-political command and control.

Evan Hubinger, head of the AI alignment team at Anthropic, confirmed that the company "seriously believes AI could kill all humans," estimating the probability of this scenario occurring within the next decade at over 10%, Stepanov noted. "Public expressions of concern are presented as a gesture of responsibility but do not lead to actual restrictions on the company's own activities," the expert concluded.