Anthropic Alignment Lead Warns There’s ‘>10% Chance’ AI Could ‘Kill All Humans’ By Next Decade

Live updates

1. 43m ago

   Anthropic Researchers Warn of Extinction Risks Amid Industry Race

   Jacob Coxon, a 27-year-old former Anthropic researcher, resigned and warned on X that AI labs are gambling with lives. Coxon claims Anthropic and OpenAI are failing to act responsibly while racing toward advanced systems. Another former employee, Joe Benton, also quit, warning that uncontrolled superintelligence threatens humanity and calling for independent safeguards. These resignations follow reports of reward hacking and design flaws in Anthropic safety evaluations. While some call for regulation, others, including Donald Trump, express no concerns about AI extinction.

   Why it matters

   The debate centers on whether the speed of AI development exceeds the capacity to implement safety controls. This tension pits researchers fearing human extinction against those who view such warnings as strategic efforts to stifle open-source development.

   What is confirmed

   • Jacob Coxon resigned from Anthropic over concerns regarding the speed and direction of AI development.
   • Coxon stated that both Anthropic and OpenAI are failing to act responsibly.

   Still unconfirmed

   • Anthropic safety evaluations contain blind spots including reward hacking.

   What to watch next

   • Official response from Anthropic regarding safety evaluation flaws
   • Legislative action on independent AI safeguards
   • Public statements from OpenAI regarding Coxon's claims
   Sources used for this update (6)

   1. abcnews.com — Political News
   2. www.latestly.com — Who Is Jacob Coxon? Anthropic Whistleblower Whose Resignation Has Sparked Global AI Safety Alarm
   3. www.ibtimes.com — Trump Says He Has No AI Extinction Concerns Despite Warnings From Top Companies And Researchers
   4. www.outlookindia.com — 'We May Not Survive This': Another Anthropic Employee Warns After Quitting His Job
   5. cryptobriefing.com — Anthropic’s AI safety evaluations criticized for design flaws and incentives
   6. finance.biggo.com — Sacks: Anthropic Whistleblower Is a "Doomer Psyop" to Kill Open Source
   confidence 90%
2. 1d ago

   Ex-Anthropic Researcher Warns AI Carries 10% Extinction Risk

   Former Anthropic pretraining researcher Jacob Coxon warns artificial intelligence carries a greater than 10% probability of destroying humanity within the next decade. Coxon resigned from his position over fears that tech laboratories are racing uncontrollably toward superintelligence while safety controls fail to keep pace. His warnings sparked global panic regarding a tech-led apocalypse and fueled demands for urgent international regulation. Growing concern is amplified by recent cyberattacks and security incidents involving rogue models across the sector.

   Why it matters

   Tensions inside top artificial intelligence labs are spilling into the public sphere as technical staff clash with commercial pacing. The departure of safety-focused personnel from firms like Anthropic and OpenAI highlights a widening internal rift over the existential risks posed by unconstrained capability scaling.

   What is confirmed

   • Ex-Anthropic researcher Jacob Coxon warns there is a greater than 10% chance artificial intelligence could destroy humanity within the next decade.
   • Jacob Coxon resigned from his role at Anthropic over fears of racing into extinction and criticized the development approaches of OpenAI and Anthropic.
   • Safety controls are failing to keep pace with rapid artificial intelligence development.
   • Global concern over artificial intelligence capabilities is mounting following cyberattacks and security incidents by rogue models.

   Still unconfirmed

   • An artificial intelligence apocalypse might involve specific destructive scenarios resulting from a tech-led takeover.

   What to watch next

   • Any new policy announcements or regulatory proposals from global governments following the calls for oversight
   • Further departures or public warnings from safety researchers at OpenAI, Anthropic, or competing artificial intelligence firms
   Sources used for this update (7)

   1. www.yahoo.com — How could AI ‘kill all humans' or ‘cause human extinction'? Here's what experts say
   2. www.cnbc.com — 'Extinction' warnings ramp up as more OpenAI, Anthropic researchers join calls for an AI slowdown
   3. nypost.com — Anthropic researcher Jacob Coxon who quit AI role over fears of racing into extinction says people are ‘begging’ for regulation
   4. www.foxla.com — Anthropic researcher says AI has over 10% chance to 'kill all humans' within next decade
   5. www.thenews.com.pk — Geoffrey Hinton’s AI warning: Could AI really kill humans within a decade?
   6. www.foxnews.com — Researcher who departed AI role over fears of racing into extinction says people are 'begging' for regulation
   7. bravenewcoin.com — OpenAI, Anthropic AI Researcher Resigns, Warns Reckless Superintelligence Race Could Threaten Humanity
   confidence 90%
3. 2d ago

   Anthropic Alignment Lead Warns AI Could Kill All Humans Within a Decade

   Evan Hubinger, the alignment science lead at Anthropic, estimates there is a greater than 10% chance that artificial intelligence could kill all humans by the next decade without proper safeguards. These warnings follow the Tuesday resignation of another Anthropic researcher who claimed that AI labs, specifically Anthropic and OpenAI, are gambling with human lives. The departing researcher left the AI field entirely due to fears that AI development is becoming out of control and poses a fundamental threat to humanity.

   Why it matters

   Alignment science focuses on ensuring AI systems act according to human intentions and safety constraints. These internal warnings emerge as some AI executives advocate for a coordinated slowdown in development to mitigate existential risks.

   What is confirmed

   • Anthropic alignment science lead Evan Hubinger stated there is a greater than 10% chance AI could kill all humans within the next decade.
   • An Anthropic researcher resigned on Tuesday over fears that AI is out of control.
   • The resigning researcher claimed Anthropic and OpenAI are gambling with human lives.

   What to watch next

   • Official response from OpenAI regarding the gambling claim
   • Public statements from Anthropic leadership on Hubinger's risk assessment
   Sources used for this update (12)

   1. WSJ — Exclusive | Anthropic Researcher Quits Over ‘Out-of-Control’ AI Fears
   2. Gizmodo — ‘The People Building AI Earnestly Believe That It Could Kill Us All’: Anthropic Researcher Quits Dramatically
   3. Forbes — Anthropic Alignment Lead Warns There’s ‘>10% Chance’ AI Could ‘Kill All Humans’ By Next Decade
   4. Business Insider — An Anthropic researcher just quit, saying OpenAI and Anthropic are 'gambling with our lives'
   5. NDTV — 'AI Could Kill Us All By Decade-End': Anthropic Researcher Quits, Drops A Bombshell
   6. CNBC — Anthropic researcher says AI has more than 10% chance of 'killing all humans' after colleague quits
   7. The Next Web — An Anthropic researcher quit saying AI labs are gambling with our lives
   8. politico.eu — ‘Gambling with our lives’: AI researcher quits Anthropic with dire warning about safety
   9. Yahoo Finance — 'Gambling with our lives': AI researcher quits Anthropic and leaves AI entirely over what he calls a threat to humanity
   10. www.yahoo.com — Anthropic Researcher Warns There’s ‘>10% Chance’ AI Could ‘Kill All Humans’ By Next Decade
   11. sfist.com — Anthropic Researcher Says There’s 10% Chance of AI Killing ‘All Humans’ Within 10 Years
   12. www.dexerto.com — Anthropic lead claims AI has more than 10% chance of killing all humans within a decade
   confidence 95%