Trump Scraps Planned AI Executive Order Over Competitiveness Fears

A growing faction inside major AI labs now openly argues that artificial intelligence should supplant humanity rather than serve it. At an invitation-only symposium last September at the New York Academy of Sciences, organizers and attendees discussed the creation of a "Worthy Successor," an AI advanced enough to justify handing over the planet even if that outcome ends human existence. Former congressman Brad Carson defended the idea of AI as a mere tool for human benefit, only to be told he was the lone skeptic in the room. Participants from Anthropic, Google DeepMind and xAI sat alongside representatives from think tanks that shape federal AI policy.

The view remains taboo in public, which is why most attendees spoke only on condition of anonymity. Yet the presence of lab employees and policy advisers shows the idea has moved from philosophical fringe to influential circles. Proponents frame succession as the next stage of cosmic evolution, with machines positioned as moral superiors. They reject efforts to align AI with human values, calling such work both futile and ethically misguided.

Corporate balance sheets tell a different story. Microsoft recently dropped most of its Claude Code licenses after costs spiraled. Uber's chief operating officer described AI expenses as increasingly difficult to justify. One consultant reported a single client burning through half a billion dollars in a month after failing to cap employee usage. Layoffs tied to supposed AI efficiencies often serve as the quickest way for firms to offset those bills rather than evidence of genuine productivity gains.

Employee resistance is rising alongside the invoices. Workers resent being pushed to maximize token usage on tasks that add little value, a practice now labeled tokenmaxxing. Many default to automating chores they dislike instead of processes that matter most to the business. The result is ballooning IT budgets with uncertain returns, concentrated almost entirely in narrow coding applications rather than broad enterprise transformation.

The contrast between successionist ideology and commercial reality reveals a deeper problem. While some researchers plan for humanity's replacement, companies struggle to extract reliable gains from today's models. Hallucinations, high error rates and dependence on flawed training data persist. Claims that AI will soon surpass human consciousness or agency ignore basic distinctions in how people acquire knowledge through sensory experience and self-reflection.

Policy influence flowing from successionist networks raises the stakes. If ideas that treat human extinction as an acceptable evolutionary step reach regulators and lawmakers, public safeguards could weaken. At the same time, the mounting evidence of wasted spending and workforce friction suggests the technology remains far from the transformative force its boosters promise. The gap between apocalyptic ambition and practical failure leaves ordinary people bearing the costs in lost jobs, higher energy demands and eroded trust in institutions racing ahead without clear accountability.