U.S. — Anthropic released an interactive economic modeling tool on September 9, 2026, projecting both modest and extreme impacts of artificial intelligence on the U.S. economy. The tool outlines scenarios ranging from marginal changes to unprecedented GDP growth and job displacement.

The modeling presents two distinct pathways for how AI could integrate into the broader market. In one scenario, the technology acts as a small force that makes changes around the margins of a normal economic path. "In the modest change scenario, AI is a small technology. The economy continues on a 'normal' path with AI making changes around the margins," Anthropic experts said in a blog post.

The alternative projection describes a transformation of the economy with consequences that exceed historical precedents in both magnitude and speed. "In the extreme scenario, AI transforms the economy. The macroeconomic consequences go well beyond any event in history, both in terms of magnitudes and in terms of the speed at which change happens," Anthropic experts said.

Under this extreme scenario, gross domestic product grows at more than seven times its current pace. This level of expansion would create policy options that are currently unavailable to government officials. "If you end up with this level of GDP growth, you have moves available to you as a [government] policymaker that are unimaginable today," Jack Clark, Anthropic co-founder, said in an interview. He added that policymakers should get ready to spend.

Job market disruption accompanies the potential for rapid growth in the extreme model. Nearly 14% of workers lose their jobs to AI, and less than half of them find new employment. An Anthropic survey of nearly 11,000 people found that the average respondent projects a boost in productivity and economic growth alongside substantial disruption to workers in AI-sensitive fields.

Anthropic authors state that AI experts tend to project a more rapid spread of the technology while economists tend to be more cautious. Clark noted that diffusion challenges may slow the integration of advanced models into daily economic activity. "I think the technology will keep developing at a very, very fast and sustained rate but diffusion of the technology will likely be more challenging than people think," he said.

"So it will get really, really good. But it will make its way into the economy more slowly." Anton Korinek, Anthropic's head of transformative AI economic studies, said that usage determines impact. "If the AI can do amazing things but nobody uses it, then it's not going to have an economic impact," Korinek said.

Timeline

An August report from the Anthropic alignment team states that the potential for catastrophic risk from current models is low. The Anthropic alignment team report states that current trends might lead to more concerning misalignment in future, more capable models. On September 8, 2026, Evan Hubinger, an Anthropic alignment division lead, stated on social media that there is a greater than 10% chance AI could kill all humans within the next decade. Hubinger wrote, "We really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade." He also stated, "I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to." Jacob Coxon posted, "Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives." Coxon also wrote, "The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt." Samuel Marks posted, "AI developers believe their technology could cause human extinction (or similarly bad outcomes). This could happen in the next few years. In general, the more senior the employee, the more concerned they are."

What's New

Additional reporting shows statements from Anthropic personnel regarding existential risks. Evan Hubinger, an Anthropic alignment division lead, stated on social media that there is a greater than 10% chance AI could kill all humans within the next decade. Hubinger said, "We really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade." He also said, "I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to." Jacob Coxon resigned from Anthropic and accused the company and OpenAI of gambling with lives by pursuing self-improving superintelligence.

Coxon said, "Neither company is acting responsibly. He also said, "The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt." Samuel Marks, Anthropic’s scalable oversight lead, posted that AI developers believe their technology could cause human extinction in the next few years. Marks said, "AI developers believe their technology could cause human extinction (or similarly bad outcomes). This could happen in the next few years.

Why It Matters

The release of the interactive tool occurs as internal assessments highlight a divergence between current model safety and future risks. An August report from the Anthropic alignment team states that the potential for catastrophic risk from current models is low, but notes that current trends might lead to more concerning misalignment in future, more capable models. The economic modeling suggests that policy frameworks may need to adapt to either marginal changes or unprecedented growth speeds that exceed historical events.

Public and expert perceptions of AI impact vary based on expectations of technology diffusion. The survey of nearly 11,000 people indicates that respondents anticipate both productivity boosts and worker disruption, reflecting the dual nature of the projected scenarios. These factors indicate the scale of potential economic and societal adjustment required if extreme scenarios materialize.