Polls show a majority of Americans now believe AI will do more harm than good, an 11-point jump in one year. This negative sentiment is growing despite, and perhaps because of, rising adoption. The paradox is that increased AI fluency correlates with decreased optimism, particularly about the job market.

In just 24 months, public perception of AI has shifted dramatically from excitement to deep concern. With Americans now five times more concerned than excited and three-quarters viewing it as a threat to humanity, the AI industry is facing a historic brand crisis rooted in fear and mistrust.

Public and expert opinions on AI are split between two extremes: it will either save humanity or destroy it. There is a notable absence of a moderate, middle-ground perspective, which is a departure from how previous technological shifts like the internet were discussed.

Anthropic's research shows that users' feelings about AI are not binary; hopes and fears coexist as tensions within individuals. The desire to use AI for learning is paired with a fear of cognitive atrophy, and the hope for productivity is tied to the fear of job displacement.

The dot-com era, despite bubble fears, was characterized by widespread public optimism. In stark contrast, the current AI boom is met with significant anxiety, with over 30% of Americans fearing AI could end humanity. This level of dread marks a fundamental shift in public sentiment toward new technology.

A significant public perception gap exists around AI. As a broad concept, "AI" is feared due to job displacement concerns. Yet, specific AI tools like chatbots are widely adopted and loved for their daily utility. This highlights a critical branding and communication challenge for the industry: the fear of the abstract versus the love of the concrete.

AI's impact has created a deep divide. Roughly half of tech workers feel amplified and excited by the technology, while the other half feel their roles have become unclear, devalued, or threatened, leading to anxiety and uncertainty.

Public opinion on AI is surprisingly negative, ranking lower than most political entities. This is driven by media focus on risks like job loss and resource consumption, overshadowing the tangible benefits experienced by millions of users. People's positive experiences with ChatGPT often coexist with a general, media-fueled distrust of "AI."

The host introduces a five-stage model describing society's relationship with AI: Skepticism, AI Psychosis, Doom Desperation, Real World Recalibration, and finally, Enlightened Excitement. This framework helps explain the volatile and often contradictory public and private reactions to AI's rapid progress.