Washington lawmakers swiftly responded to growing anxieties over advanced artificial intelligence capabilities. Representative Lori Trahan, a Democrat representing Massachusetts, noted that key safety personnel are stepping down while powerful computer systems exceed lab boundaries despite ongoing corporate momentum, highlighting bipartisan proposals like the FRONTIER Act and the Ban Artificial Superintelligence Act. Meanwhile, Senator Bernie Sanders pledged to introduce measures halting technological expansion, and Senator Ted Cruz characterized the existential threat as “scary as hell”. Furthermore, David Sacks, who chairs the President’s Council of Advisors on Science and Technology, demanded a halt to Anthropic’s planned initial public offering—slated for mid-October at the earliest prior to the November midterms—until a thorough review of the whistleblower allegations occurs.
This political momentum stems from viral disclosures made by former Anthropic staffer Jacob Coxon following his departure from the organization. Having logged roughly three years of research split between OpenAI and Anthropic, Coxon asserted on social media that practitioners in the field genuinely anticipate these systems could eliminate humanity prior to 2030. He declared that technology firms were “racing straight to self-improving superintelligence and gambling with our lives”, a message that generated over 150 million impressions according to Bloomberg and topped 100 million in merely a few days via Wired. Speaking with Axios, Coxon revealed he walked away after approximately four months at Anthropic, foregoing his equity vest by two months to ensure he derived no commercial benefit from boosting the startup’s financial valuation.
Prominent engineers across leading organizations validated these apocalyptic warnings, including Evan Hubinger, head of Anthropic’s alignment science division, who estimated a greater than 10 percent probability of human demise within ten years. His colleague Samuel Marks observed that concern escalates alongside job seniority, while former U.S. Commerce Department safety chief Paul Christiano cautioned that rapid progress risks an irrevocable, catastrophic breakdown in control where most citizens perish, shortly before OpenAI appointed him to its non-profit board. Additional voices chimed in, with pioneer Geoffrey Hinton telling BBC Newsnight that a 10 percent extinction odds calculation was reasonable, alongside OpenAI’s Julie Steele advocating for an industry slowdown in her personal capacity. These public statements reinforce a July petition endorsed by nearly 1,400 industry personnel demanding federal oversight to moderate automated AI advancement.
Conversely, military leadership actively dismissed the apocalyptic narratives during a Washington defense gathering on Thursday. Emil Michael, chief technology officer for the Department of Defense, downplayed the viral claims as an overhyped psychological loop, arguing that commercial competition, collective industry efforts, and state collaboration would successfully address potential damages. Michael further revealed the Pentagon had already transitioned roughly 90 percent of its secret computational tasks away from Anthropic amid an ongoing legal battle over military deployment boundaries, with the final phase completing before month’s end. Furthermore, independent expert Gary Marcus informed Business Insider that while digital assaults and deceptive media present genuine threats, he sees no plausible mechanism through which AI exterminates humanity.
Neither Anthropic nor OpenAI offered immediate comments to the Associated Press regarding Coxon’s resignation, though an Anthropic representative told CNBC that the enterprise maintains robust safeguards and was the pioneering laboratory to issue a framework for curbing severe systemic hazards. The company emphasized its long-standing openness regarding both the extraordinary rewards and unprecedented hazards of synthetic intelligence. While figures such as Turing Award winner Yoshua Bengio, former Google scholar Geoffrey Hinton, and OpenAI chief scientist Jakub Pachocki have previously urged extreme care, the sheer breadth of this week’s public engagement and high-level scientist endorsements marked an unprecedented turning point in the global AI safety debate.
Leave a Reply