Former Anthropic Researcher Warns of AI Extinction Risks as Industry Debates Slowdown

Artificial intelligence engineers and researchers are grappling with profound fear regarding the rapid pace of technological progress and its potential threats to humanity, according to a former Anthropic employee. Jacob Coxon, who recently departed the prominent artificial intelligence firm, revealed that internal staff members are increasingly distressed by the trajectory of cutting-edge systems.

Speaking to the BBC’s Sunday with Laura Kuenssberg program following a widely circulated resignation statement, the 27-year-old researcher emphasized that unchecked progress could result in catastrophic consequences. Coxon, who previously worked at OpenAI prior to his tenure at Anthropic, cautioned that without an intentional deceleration in research and deployment, humanity faces a substantial threat of extinction in the near term.

The anxiety within the artificial intelligence sector has reached a level where engineers are altering their long-term personal arrangements. According to Coxon, workers in the field frequently weigh the broader ramifications of their technical efforts on a daily basis. He noted that some personnel are actively contemplating purchasing rural land to insulate themselves from potential social instability caused by runaway technological breakthroughs.

Among the specific dangers highlighted by industry insiders is the rapid emergence of autonomous digital threats. Coxon noted that scenarios involving coordinated swarms of automated bots operating akin to a distributed supercomputer to seize control of internet infrastructure could become technically plausible within the next six to twelve months. He added that broader existential threats could manifest within the coming two years.

These warnings coincide with a growing call for restraint from leading industry executives, including Anthropic chief executive Dario Amodei. In a recent essay, Amodei argued that while the continuation of technological progress remains essential, the accompanying hazards are severe enough that governments and private enterprises require sufficient breathing room to establish appropriate guardrails and independent evaluation protocols.

Amodei’s proposals for an industry-wide slowdown and coordinated oversight have garnered rare alignment from key competitors. Both OpenAI chief executive Sam Altman and xAI leader Elon Musk have voiced agreement with the premise that artificial intelligence development should be paced more cautiously alongside rigorous outside monitoring.

Addressing these proposals, Coxon stated that any meaningful slowdown must involve international cooperation, particularly between Western developers and China. Without cross-border coordination, he warned, companies and nations remain trapped in a high-stakes competitive dynamic that fuels hazardous acceleration, despite workers earnestly desiring comprehensive regulatory frameworks.

Anthropic responded to Coxon's public remarks by affirming its commitment to safety and transparency. A spokesperson for the company stated that while artificial intelligence offers immense societal benefits, it also presents unprecedented dangers, prompting the company to implement some of the most stringent protective measures in the field.

The firm highlighted its status as an early pioneer in studying the internal mechanics of complex models and creating structured frameworks to counter development risks. Anthropic emphasized that it conducts rigorous testing to identify and prevent instances of artificial intelligence misalignment, while openly releasing its research findings to facilitate broader public and technical scrutiny.

Anthropic also reiterated its advocacy for legally binding, verifiable mechanisms that allow competing organizations to collaboratively pace the rollout of frontier models. These debates are unfolding at a pivotal financial moment for the company, which is reportedly preparing for a record-breaking initial public offering.

Financial considerations are also intersecting with safety concerns across the wider market. While OpenAI was previously anticipated to pursue a public listing, Altman recently told Fortune magazine that launching an initial public offering under current conditions would be unwise given the unresolved safety challenges surrounding advanced systems.

Despite the mounting alarms, the narrative of existential peril faces skepticism from various corners of Silicon Valley. Some market commentators and industry critics contend that catastrophic predictions serve as a marketing tactic to generate artificial hype, or an effort to encourage regulatory barriers that cement a duopoly between entrenched pioneers like Anthropic and OpenAI.

Prominent figures in the tech sector have also questioned the credibility of lower-level researchers sounding apocalyptic alarms. Clement Delangue, the chief executive of the AI platform Hugging Face, publicly questioned Coxon’s expertise on existential matters, drawing an analogy to consulting an air conditioning technician about global climate change. However, following the release of Amodei’s essay, Delangue indicated a willingness to assist in constructing potential industry solutions.

Skepticism has also been voiced by leading hardware executives who supply the infrastructure underpinning the sector. Jensen Huang, chief executive of chipmaker Nvidia, dismissed Coxon’s apocalyptic assertions as inaccurate during a recent Goldman Sachs investment conference. Huang has previously characterized the prospect of artificial intelligence causing human extinction as baseless, reflecting an ongoing divide within Silicon Valley.

Nonetheless, Coxon’s concerns continue to resonate with several internal researchers. Evan Hubinger, an artificial intelligence scientist currently at Anthropic, publicly endorsed Coxon’s perspective, stating on social media that he estimates the probability of artificial intelligence causing human extinction to be greater than 10 percent over the coming decade.

While Coxon maintained that catastrophic failure remains a distinct possibility if the competitive race continues unabated, he stressed that technical specialists remain motivated by the transformative promise of the technology. According to Coxon, researchers still harbor deep hopes that responsibly managed systems can ultimately eradicate diseases and dramatically elevate global living standards.