Why people really hate AI

What It Covers

David Pierce and Nilay Patel analyze why public sentiment toward AI remains deeply negative despite widespread adoption, examining OpenAI's internal pivot memo from CEO of Applications Fidji Simo, NBC polling showing 57% of Americans view AI risks as outweighing benefits, and why the industry has failed to produce a mainstream consumer product comparable to the iPhone or Instagram.

Key Questions Answered

• •Consumer AI monetization failure: OpenAI's Fidji Simo issued an internal memo declaring a pivot away from broad consumer products toward enterprise and coding use cases — the only segment showing genuine product-market fit. ChatGPT attracts massive usage but loses money on every interaction, and attempts at ads and shopping have not moved toward profitability. The memo signals no viable path exists from consumer scale to consumer revenue without a fundamental business model shift.
• •The 57% problem: An NBC News poll shows 57% of Americans believe AI risks outweigh benefits, versus only 34% who disagree. A separate Pew Research study found 53% believe AI will worsen creative thinking, and 50% believe it will damage meaningful human relationships. Only 5% said AI would improve relationships. These numbers have remained consistently negative throughout AI's rise — this is not a recent backlash but a sustained baseline of public skepticism.
• •Environmental objections don't move consumers: VC arguments that negative media coverage about AI's water and energy consumption is driving public hostility misread how consumers actually behave. Fast fashion, oversized gas vehicles, and single-use plastics all demonstrate that Americans consistently choose convenience over environmental concern. Blaming water-usage headlines for AI's unpopularity ignores decades of evidence that environmental messaging alone has never meaningfully shifted mainstream consumer purchasing behavior.
• •The "social permission" threshold: Microsoft CEO Satya Nadella stated publicly at Davos that AI companies risk losing social permission to consume scarce resources like energy unless the technology demonstrably improves health outcomes, education outcomes, and public sector efficiency. This framing identifies the core problem: the industry is extracting enormous public resources — land, power, RAM, data — without delivering consumer value that justifies the extraction in the minds of ordinary people.
• •Killer app benchmark — Uber and Instagram: The mobile revolution succeeded because it produced use cases impossible without the hardware: Uber put a car in anyone's hand via one button press, and Instagram placed distribution directly next to the camera, collapsing the gap between creation and audience. AI has produced no equivalent. Coding tools and enterprise software are real but not mainstream. Vibe coding and personal automation are not behaviors most people identify with or want to adopt.
• •VC messaging created the backlash: For years, AI investors and founders including Sam Altman publicly predicted mass job elimination and economic restructuring, framing this as inevitable and positive. When consumers reacted negatively to the prospect of losing their jobs, the same investors began labeling that reaction "doomerism" and blaming founders like Dario Amodei. The hosts argue this is circular: the industry generated the fear through its own fundraising narrative and now deflects responsibility for the resulting public hostility.
• •Enterprise is the real AI business: The clearest evidence of AI's actual value sits in B2B software — coding assistants, SaaS workflow tools, and developer productivity. Companies built from the ground up on AI agent cost structures can undercut legacy software vendors. However, this disruption path does not scale to the consumer revenue needed to justify current data center investment levels. The industry's existential challenge is that enterprise success alone cannot validate the trillion-dollar infrastructure buildout already underway.