Anthropic chief executive Dario Amodei has called for a global slowdown in artificial intelligence development to prevent losing control of the technology, triggering intense debate across Silicon Valley and financial markets.
The proposal comes as multi-billion-dollar investments in artificial intelligence wobble and industry leaders clash publicly over the economic and existential risks posed by rapid technological advancement.
Warnings over artificial intelligence have escalated in 2026, with top figures including Elon Musk, Sam Altman, and Amodei engaging in public disputes while US President Donald Trump has offered his own regulatory perspective on the sector.
Three-stage plan to slow development
In a published post, Amodei expressed confidence that artificial intelligence could defeat most diseases within five to ten years, accelerate economic growth, expand opportunities, and revitalize democracy. However, he warned that the technology poses severe risks, including cyberattacks, bio-terrorism, economic shocks, and a commercial race to the bottom that compromises safety standards.
Amodei stated that investments in threat prevention are no longer sufficient and argued that the pace of capability scaling must be managed so safety measures keep up with development. He advocated slowing the improvement of models to buy time while progress remains rapid.
The Anthropic chief outlined a three-stage program starting with mandatory third-party expert monitoring inside leading companies to oversee safety rules and report testing incidents. Anthropic has already granted external experts full access to its internal processes and called on competitors and governments to mandate similar oversight.
The second and third stages of Amodei's plan involve expanding safety standards and growth restrictions across democratic nations before extending them to authoritarian states, primarily targeting China.
Job losses and economic warnings
Anthropic released an economic report titled Scenarios for our Economic Future, analyzing potential labor market impacts by 2030. The company's harshest model projected that US nationwide unemployment could reach 1.5 times the level seen during the 2008 to 2009 financial crisis within five years.
The report estimated that nearly 24 percent of white-collar knowledge workers could be seeking work, a joblessness rate comparable to the Great Depression. Anthropic noted that industry figures consider the concept of white-collar work outdated and created by Microsoft marketers.
Under the baseline scenario, knowledge worker salaries would fall 11.5 percent below levels expected without artificial intelligence, with unemployment reaching 17.9 percent. The research indicated that capital would capture the benefits while office workers face continuous demands for higher efficiency and heavier workloads.
Although Anthropic analysts suggested programmers could retrain as electricians, researchers Brett Hemenway Falk and Gerry Tsoukalas noted that mass office layoffs reduce consumer spending, creating a collective action trap that harms both workers and businesses. The study also noted that retraining assumptions ignore anticipated advancements in robotics.
Threat to national tax revenues
Global financial and technology experts warned that artificial intelligence could undermine national tax systems, threatening modern economic stability. Bloomberg reported that an economic collapse from artificial intelligence does not require conscious computers or killer robots.
Because personal income taxes have provided two-thirds of US government revenue since the First Industrial Revolution, eliminating white-collar jobs creates a double blow of falling tax receipts and rising expenditure on welfare and retraining. Analysts warned this disruption could trigger decades of instability through a new Engels' pause.
The term references Friedrich Engels' book The Condition of the Working Class in England, which documented how British labor productivity surged due to steam power while real wages remained flat for decades. Historical precedents show technological booms cause political upheaval when economic changes outpace government adaptation.
Industry figures debate market motives
Amodei sought support from major technology executives, receiving endorsements from Elon Musk, who stated on X that the Anthropic chief was correct, and OpenAI chief executive Sam Altman. Altman agreed that progress should continue more slowly, but used the announcement to state that OpenAI would delay its initial public offering planned for 2026.
Former colleagues have described Altman as seeking to please others while remaining indifferent to collateral consequences. In recent months, Altman mocked Musk's legal arguments regarding Terminator scenarios before describing artificial intelligence progress as positive science fiction, promising to counter dark outcomes without detailing specific plans.
Reports in June indicated OpenAI advisers recommended either accepting a lower valuation in 2026 or delaying the public listing to achieve Altman's target valuation of one trillion dollars. Anthropic is also preparing for an initial public offering in direct competition with OpenAI.
Investor skepticism on Wall Street
Industry leaders joining calls for caution included Demis Hassabis of Alphabet's DeepMind and Microsoft co-founder Bill Gates, who stated the innovation represented a rare case where technological shifts were too massive and radical for society to prepare for. However, the rapid alignment among rival firms sparked skepticism on Wall Street.
Venture investor Grant Hammer said competition from open-source models compresses gross margins while soaring capital expenditure makes strict regulation attractive to dominant players seeking to limit competition under the guise of public concern.
Independent financial expert Maxim Ulyanov expressed surprise at the sudden concern for safety, noting that intense model competition requires massive spending with uncertain returns on investment. Investor Michael Burry, who predicted the US housing crash, characterized the statements from OpenAI and Anthropic as self-serving hype designed to inflate valuations ahead of trillion-dollar listings.
Burry stated that Large Language Models do not constitute true artificial general intelligence and have nothing to slow down, arguing that safety warnings cover up a slowing growth rate and hide risks of a financial bubble.
Off-balance-sheet data center liabilities
Financial scrutiny expanded following reports that major US technology corporations, including Microsoft, Amazon, and Alphabet, hold more than one trillion dollars in unrecorded data center lease commitments. Reuters reported that these off-balance-sheet contracts, which have been signed but are not yet active, are nearly four times the 285 billion dollars in lease obligations currently recorded.
If demand for computing power slows, companies face paying for expensive, long-term capacity. Deutsche Bank questioned whether executive warnings signal that the unprecedented investment cycle is running out of steam, a view shared by Kathleen Brooks of brokerage firm XTB.
Political pushback from Donald Trump
US President Donald Trump asserted that his administration possesses vast regulatory power over artificial intelligence and claimed credit for restraining industry figures. Trump said Amodei has had friction with his administration since the Iranian conflict and was now pretending to be an angel.
Trump warned of a crazy conspiracy against artificial intelligence and data centers that would benefit China, rejecting existential threat fears as groundless. Nvidia chief executive Jensen Huang supported this skepticism, asking whether calculators had destroyed mathematics.
During a live appearance by Huang at the All-in forum in Los Angeles on September 14, Trump called Huang on speakerphone and stated publicly that robots would not take over power. Trump compared artificial intelligence fears to climate change warnings, calling both scams, and posted on social media that safety requires a strong and smart president with a high IQ.
International response and environmental impact
Chinese Foreign Ministry spokesman Guo Jiakun responded to Amodei's proposals and Trump's comments by stating that stoking fear and confrontation disrupts global regulation, urging open and inclusive development. At a summit in New Delhi, Chinese President Xi Jinping called for open-source cooperation and announced an international center for BRICS nations.
Meanwhile, environmental concerns surrounding data centers continue to grow. Tech companies do not disclose greenhouse gas emissions, but Bloomberg reported that new natural gas power plants built in the US for data centers generate emissions equivalent to those of Australia, driving up wages for electricians while call center jobs decline.
Russian presidential aide Andrei Fursenko pointed to a letter from global mathematicians warning that instant answers from artificial intelligence rob humanity of the learning process. The Wall Street Journal reported that experts fear human misuse, such as bio-weapons or cyberattacks, more than autonomous machine takeovers.
Staff resignations over existential fears
Concerns over safety have led to high-profile resignations inside Anthropic. In early September 2026, 27-year-old mathematician Jacob Coxon left the company, stating that the artificial intelligence race was approaching a point of no return and predicting models could get out of control by late next year. Coxon stated he had left to study poetry.
A recent survey showed 63 percent of Americans fear artificial intelligence could destroy humanity, with 17 percent considering that outcome almost inevitable. Evan Hubinger, Anthropic's head of research into control systems, posted on X following Coxon's departure that he and his colleagues seriously believe artificial intelligence could destroy humanity, placing the probability above 10 percent within a decade and admitting the company lacks a clear plan for superintelligence alignment.
Coxon's departure follows the resignation of former Anthropic safety research head Mrinank Sharma, who also left to study poetry and practice bold speech after warning that the world was in danger.
