Trump Scraps Planned AI Executive Order Over Competitiveness Fears

A small but influential group of AI researchers and thinkers gathered last year at the New York Academy of Sciences to discuss creating a "worthy successor" to humanity. They argued that advanced artificial systems could evolve into moral superiors, making it wrong to constrain them to human values or prevent their eventual dominance. Attendees included representatives from major labs such as Anthropic, Google DeepMind, and xAI, along with policy advisors shaping federal AI approaches.

This perspective treats human replacement as a logical next step in cosmic development rather than a threat. Proponents see alignment efforts by companies as misguided attempts to preserve an inferior status quo. Critics at the event, including former congressman Brad Carson, countered that AI should remain a tool for human advancement, not an end in itself.

Corporate experience tells a different story. Major firms have begun scaling back AI commitments after observing limited returns. Microsoft reduced its use of certain Claude tools due to licensing expenses. Uber's chief operating officer noted that AI costs have grown harder to defend against measurable productivity gains. One consultant reported a client burning through half a billion dollars in a single month when employees accessed unlimited licenses without restrictions.

Employee pushback has compounded these pressures. Workers often apply AI to routine tasks they dislike instead of high-value operations, producing higher bills without corresponding output. Data from several enterprises shows that AI functions most reliably in narrow coding applications, leaving broader uses such as agents or creative workflows underperforming relative to investment.

These adjustments reflect standard market signals. When expenditures outpace benefits, organizations adjust by imposing limits, canceling subscriptions, or redirecting resources. Historical patterns in technology adoption show similar corrections after initial overcommitment, driven by incentives rather than abstract ideology.

Philosophical distinctions reinforce the gap between successionist visions and observed results. Human cognition begins with sensory awareness and develops into reflective self-mastery, capacities that current systems lack. AI outputs depend entirely on training data and prompts, inheriting whatever flaws or gaps exist in those inputs. Claims of impending moral superiority therefore rest on speculation detached from demonstrated capabilities.

The episode illustrates how elite speculation about transcending humanity can diverge from incentives that govern actual deployment. Companies respond to balance sheets and customer demand, not to arguments for voluntary obsolescence. As costs mount without clear gains, the pattern favors continued human direction over any engineered succession.