80,000 Users confess to claude: we want ai to steal the boring bits of our jobs

Anthropic just turned the loudest AI ethics debate into raw numbers. In December the lab let its own model interview 80,508 people across 159 countries, in 70 languages, with one sneaky prompt: “If you had a magic wand, what would AI do for you?” The answer is less sci-fi than the valley expected, and more mercenary.

Professional excellence tops the chart at 18.8 %. Not world peace, not robot lovers—just the mundane promise of inbox zero and slide decks that build themselves. The community wants AI as a private intern, not a messiah.

The hidden hierarchy of human desire

Strip away the hype and the rest of the ranking reads like a diary of late-capitalist fatigue. Personal transformation (13.7 %) and life management (13.5 %) come next, hinting that even yoga apps have lost their shine. People crave an algorithmic nudge that books the dentist, remembers birthdays and still leaves room for self-actualisation.

Time freedom lags just behind at 11.1 %, a figure that should chill every gig-platform executive still selling 24/7 hustle. The same respondents who chase excellence also long to escape the hamster wheel; the contradiction is the point.

Why the global south skews the data

Anthropic’s heat map reveals a stark divide. Users in the US and Western Europe talk about creative expression and learning; respondents in Nigeria, Brazil and Indonesia talk about financial independence and entrepreneurship. The model, trained mostly on English corpora, still detected divergent grammar of aspiration. One Spanish-speaking student asked the AI to “make my mother retire early”; a coder in Lagos wanted it to “find the cheapest flight so I can pitch in Silicon Valley.”

The lab refused to weight answers by GDP, so the 9.7 % who chose “financial independence” punch above their economic footprint. They are not dreaming of yachts; they want breathing room.

What anthropic refuses to spell out

Here’s the rub: the survey ran inside Claude’s chat interface, meaning participants already self-selected as AI-curious. The 80,508 voices exclude the half-planet still suspicious of anything with a transformer backbone. Yet even inside the echo chamber, only 5.6 % care about “creative expression,” the use case most startups brandish in pitch decks.

Translation: the market is speaking, and it is asking for boring automation, not generative poetry. Investors pouring cash into AI-powered social networks might want to re-read the spreadsheet.

The fine print that will shape 2026 policy

European regulators scanning the study will notice that “social transformation” sits at 9.4 %, just below money worries. That cohort wants bias audits, explainability, maybe even algorithmic reparations. Anthropic handed them a free lobbying document: citizens want the tech optimised first for dignity, second for disruption.

Meanwhile, 8.4 % still voted for “learning and growth,” the category closest to education ministries’ talking points. EdTech vendors will weaponise that sliver in every Brussels hearing, claiming their chat tutors align with public demand. The numbers are small, but politics runs on anecdotes dressed as data.

Bottom line

Strip the survey to its studs and you get a single, blunt request: make the grind bearable. We don’t want wonder; we want a quieter Monday. If Anthropic’s competitors ignore that memo, someone else will write the code—and cash the cheque.