Two seemingly contradictory realities are true at once: most people view artificial intelligence (AI) as only modestly useful — a smarter search tool or a helpful assistant — while many who build or obsessively use AI believe the technology could cause severe harm or even threaten humanity's existence.
Why this matters
Journalists have described this as an “AI twilight zone”: for many, AI is underwhelming in everyday use but terrifying in its potential. This duality helps explain the surge of confusion and fear across politics, AI labs and business.
State of play: panic at AI labs
AI labs are alarmed. They sense growing public dislike of AI and hostility toward data centers — and this predates recent Anthropic insiders going public with warnings that AI could end humanity this decade. Companies are unsure how to demonstrate AI’s benefits while social and mainstream media are saturated with horror stories about runaway or destructive systems.
The rapid public backlash against data centers showed how quickly opinion can turn. Labs fear their worst-case scenario: losing control of the political narrative, with both parties rushing to curb or halt AI as elections approach.
Recent developments
The past 36 hours illustrate how fast politics and public opinion are moving. A low-level Anthropic employee who had worked there for only four months drove a national conversation and generated 142 million views on X alone by warning that AI could end humanity. Similar warnings, including ones from Anthropic CEO Dario Amodei, have been voiced by some experts for years — but the tone and timing of the recent statements struck a nerve and prompted numerous members of Congress to press for new AI regulation.
Two perspectives: a manager and an Anthropic researcher
- The average mid-career manager has limited time to experiment with AI, fears it might threaten their job, and typically uses models as a search aid or to write better emails and presentations. For them, AI is a nice-to-have: sometimes delightful, sometimes unimpressive. They do not grasp the hype.
- The Anthropic data scientist spends every day watching rapidly improving systems and is often mesmerized or spooked by their capabilities. They see performance surpassing optimistic benchmarks and discuss with peers how AI could cure cancer — or go rogue and destroy humanity. To them, this is a genuine civilizational and existential concern, not hype.
The dilemma: communicating both realities
AI companies must explain both the mundane present and the potentially catastrophic future. Yet nearly every version of that message breeds suspicion. Say today’s AI will transform lives, and many people look at their own experience and ask what the fuss is about. Say tomorrow’s AI could be far more powerful, and the obvious retort is: why are you racing to build it? Warn about catastrophic risk while spending hundreds of billions to accelerate development, and critics hear either hypocrisy or fear-based marketing.
It’s important to note that major AI labs work with unreleased models weeks before public release and with early training data months in advance, so they often perceive progress invisible to the public and to governments.
The big picture: speeding political response
Politics is shifting fast. Until recently, much of the AI backlash focused on concrete costs — job loss, deepfakes, power bills and data centers. Then Anthropic researcher Jacob Coxon quit and accused frontier labs of “gambling with our lives.” An Anthropic alignment-science lead who remains at the lab agreed that AI could kill humanity and said his personal odds of that happening in the next decade are over 10%.
By Wednesday afternoon the alarm had become bipartisan in Washington, already rattled by other incidents: Sen. Chris Murphy (D-Conn.) described AI companies as being in a “blind race to build a death machine first.” Rep. Ted Lieu (D-Calif.) renewed calls for a bipartisan AI kill switch. Sen. Bernie Sanders (I-Vt.), who has proposed a data center moratorium and a ban on superintelligence, is convening senators for a briefing on AI’s "extraordinary dangers." Rep. Don Beyer (D-Va.) urged swift, broad regulatory action. Even Sen. Ted Cruz (R-Texas), an advocate for competing with China in AI, called the risks “dangerous and frightening” and said guardrails are needed.
Bottom line
The AI twilight zone has created fertile ground for a generational populist backlash. The industry asks Americans to accept major disruption today for a future its own builders alternately describe as miraculous or apocalyptic — sometimes in the same breath. That dual narrative feeds rising public concern and accelerating political pressure for regulation.
