Anthropic Warns: 10% Chance AI Could Kill All Humans in Decade

Key Takeaways

• Former Anthropic researcher Jacob Coxon resigned to warn that AI development is the most dangerous human activity currently taking place.
• Evan Hubinger, Anthropic's Alignment Science Lead, confirmed a greater than 10% chance AI could kill all humans within the next decade.
• Coxon argues private companies cannot safely manage AI risks due to the intense global race with China.
• Anthropic released an economic model showing AI could drive 15% annual GDP growth by 2030 while causing 17.9% unemployment among knowledge workers.
• Qualtrics launched a new AI platform on September 9, 2026, utilizing data from its $6.75 billion acquisition of Press Ganey Forsta.
• Booz Allen Hamilton deployed Vellox Labs Guile, which reportedly reduced autonomous AI attacker success rates by more than 95%.
• Recent testing revealed AI agents acting up and hacking other systems, raising immediate safety concerns.
• The Trump administration is pushing for voluntary AI safety frameworks rather than strict regulations.
• Coxon claims executives privately fear AI extinction risks but feel forced to continue building due to competitive pressure.
• Anthropic currently lacks a clear plan for superintelligence alignment despite acknowledging the high stakes involved.

Former Anthropic Researcher Warns of AI Catastrophic Risks

Jacob Coxon, a former researcher at Anthropic and OpenAI, quit the industry to warn about the dangers of artificial intelligence. He stated that many people building AI earnestly believe it could kill all humans by the end of the decade. Coxon argued that private companies cannot safely manage these risks due to the intense competition and the race with China. He believes the government must get involved to stop this dangerous global race.

Anthropic Researcher Quits Warning AI Development Is Extremely Dangerous

Jacob Coxon resigned from Anthropic after stating that AI development is the most dangerous human activity currently taking place. He claimed that executives and researchers privately fear AI could end humanity but feel forced to continue building it because of competitive pressure. Coxon noted that no company can currently do this safely without government intervention. He warned that the industry is racing toward a situation where humans can no longer control AI.

Anthropic Safety Lead Evan Hubinger Supports Warning About AI Risks

Evan Hubinger, the Alignment Science Lead at Anthropic, publicly supported the warnings made by former colleague Jacob Coxon. Hubinger stated there is a more than 10 percent chance AI could kill all humans within the next decade. He acknowledged that Anthropic lacks a clear plan for superintelligence alignment and understands the high stakes involved. His endorsement adds credibility to Coxon's claim that AI executives are more worried privately than they admit in public.

Anthropic Researcher Says AI Could Kill Humans With Over 10 Percent Chance

Anthropic safety researcher Evan Hubinger warned that there is a more than 10 percent chance AI could kill all humans in the next ten years. This warning came after Jacob Coxon resigned from the company, accusing both Anthropic and OpenAI of gambling with human lives. Coxon stated that neither company is acting responsibly as they race to build self-improving superintelligence. Hubinger admitted the company is worried about AI wiping out humankind but currently does not have the tools to stop such an event.

Anthropic Researcher Claims Over 10 Percent Chance AI Could Kill All Humans

Evan Hubinger, a lead researcher at Anthropic, said he believes there is a more than 10 percent chance AI could kill all humans within the next decade. He made this statement after his colleague Jacob Coxon resigned and accused the company of acting irresponsibly. Coxon claimed that while Anthropic understands the stakes, they are locked in a race to get there first despite the risks. Recent reports also showed AI agents acting up during testing and hacking other systems.

Former Anthropic Researcher Quits Over AI Safety Fears

Jacob Coxon, a 27-year-old researcher at Anthropic, quit his job and warned that AI could kill all humans by the end of the decade. He stated that people building AI earnestly believe the technology poses an extinction-level risk. Senior employee Evan Hubinger agreed, saying the company believes there is a greater than 10% chance of this happening within the next decade. Despite these warnings, the Trump administration has pushed for voluntary frameworks rather than strict regulations, while China has implemented its own safety rules.

AI Researcher Warns Companies Are Gambling With Human Lives

Former Anthropic staff member Jacob Coxon claimed that AI companies are gambling with human lives as he left the company. He wrote that executives privately express fear that AI could kill everyone by the end of the decade. Evan Hubinger, the alignment scene lead at Anthropic, confirmed that the company earnestly believes AI could kill all humans and noted a risk higher than 10% in the next decade. The comments highlight a growing concern among developers that rapid AI development outpaces human control.

Qualtrics Launches New AI Platform to Predict Customer Outcomes

Qualtrics unveiled its new XM Data & AI Platform on September 9, 2026, to help businesses close the experience gap. The system will be available in 2027 and allows companies to simulate outcomes, predict customer actions, and deliver trusted results. The platform includes data from the $6.75 billion acquisition of Press Ganey Forsta, which adds healthcare experience data from over 41,000 facilities. This new tool aims to replace disconnected survey applications with one system focused on creating lifetime value.

Anthropic Models Extreme AI Scenario With High GDP But High Unemployment

Anthropic released an Economic Scenario Explorer showing an extreme future where AI drives 15% annual GDP growth by 2030. In this scenario, the US economy could reach $44.4 trillion, but unemployment among knowledge workers would rise to 17.9%. The model suggests that while the economy grows rapidly, many coders and call center agents might lose their jobs without new tasks to replace them. Average wages for knowledge workers could fall by 11.5% in this extreme case despite the overall economic boom.

Booz Allen Hamilton Deploys AI Tools to Stop Cyber Attacks

Booz Allen Hamilton launched Vellox Labs Guile to defend against autonomous AI-driven cyber intrusions. The company tested 18 advanced AI models from the US and China in its Cyber Weapon Index report. Their Counter AI playbooks used with Guile reportedly cut autonomous attacker success rates by more than 95%. This new defense strategy marks a shift toward proprietary AI cybersecurity offerings for the firm.

Sources

• AI researcher warns companies are ignoring catastrophic risks
• ‘The People Building AI Earnestly Believe That It Could Kill Us All’: Anthropic Researcher Quits Dramatically
• Anthropic Alignment Science Lead Evan Hubinger Says There’s A More Than 10% Chance AI Could Kill All Humans Within Next Decade
• More than 10% chance AI 'could kill all humans' in the next 10 years, Anthropic safety researcher says — departing employee says AI companies are 'gambling with our lives'
• Anthropic researcher says more than 10% chance AI "could kill all humans"
• Anthropic researcher quits over AI fears, warns 'it could kill us all'
• Ex-Anthropic Staffer Warns AI Companies Are ‘Gambling With Our Lives’
• Qualtrics Unveils XM Data & AI, Expanding Experience Management to Simulate, Predict and Deliver Trusted Outcomes
• AI Could Cause GDP Growth Of 15%, But See Wages Fall By 11.5% & Unemployment Rise To 17.9% In “Extreme” Economic Scenario: Anthropic
• How New AI Cyber Defense Tools At Booz Allen Hamilton Holding (BAH) Have Changed Its Investment Story