Who Actually Pays for AI in China? The Hidden-Bill Essay Going Around

A July 6, 2026 essay circulating on Chinese social media argues that AI’s costs are quietly shifting onto ordinary workers and consumers. It opens with a company that replaced 3 customer-service staff with a chatbot to save ¥360,000 a year, cites the World Economic Forum’s Future of Jobs Report 2025 (85 million jobs displaced, 97 million created globally), and claims the losses hit low-wage roles while the gains go to high-skill ones — and that users of “free” AI tools are paying with their data.

The numbers

• 3 customer-service employees replaced by a chatbot; ¥360,000 saved in one year [as reported in the essay]
• 85 million jobs displaced / 97 million created globally (WEF Future of Jobs Report 2025 — sourced)
• ¥5,000/month typical salary of disappearing roles [unverified]
• $187,000 median salary for “AI ethics advisor” [unverified]
• 60% of firms cut hiring or laid off staff due to AI’s potential (attributed to Harvard Business Review, 2025) [unverified]
• Over 60% of orders on major Chinese hiring platforms first screened by AI [unverified]
• AI fortune-telling tools: a few yuan to tens of yuan per reading; ¥20 example “life guidance” reading
• Example subscription stack: ¥200/month on copywriting, slides, translation tools
• Freelancer case: output doubled from 3 to 6 articles/day (self-publishing platform) — income did not follow
• AI task complexity doubles every 4–7 months (attributed to the UN’s first global AI assessment) [unverified]

The take inside China

The essay’s punch is simple: AI’s costs are being paid by people who never agreed to pay them. The ¥360,000 a company saves doesn’t reach the laid-off workers or the customers — per the essay it flows to tech-giant subscription fees, shareholder dividends, and executive bonuses [unverified].

The sharpest point is the “free” trap. ChatGPT costs nothing because you’re not the user — you’re the training data. And who feeds free AI the most? Junior copywriters, data analysts, translators, customer-service staff: exactly the people AI can most easily replace. Meanwhile, over 60% of hiring-platform orders reportedly start with an AI filter that reads your projects, client reviews, even your reply speed [unverified].

Then there’s 2026’s boom in AI fortune-telling — a few yuan for a “prediction,” paid in birth dates, relationship status, career anxieties. The essay’s verdict: not fortune-telling, but user profiling in disguise — buying your psychological weak points for ¥20.

For freelancers, the “efficiency trap”: a copywriter doubled output from 3 to 6 pieces a day; income didn’t follow, because platforms cut per-piece rates when everyone speeds up. Efficiency became table stakes, not leverage.

The essay’s advice: use AI to replace paid services, not add subscriptions; use it to learn skills, not kill time; guard your data. And Hinton’s second half — AI creates new possibilities “if people are prepared” — is framed as the real skill of the AI era.

Why it matters outside

This is a mainstream articulation of a view increasingly common in China: AI’s benefits accrue to capital and platforms while its costs land on low-wage workers and data-giving users — dynamics Western readers will recognize from their own debates. Expect this “who pays for AI” framing to shape Chinese public opinion on regulation, labor policy, and platform trust.

Sources

• WEF Future of Jobs Report 2025

Provenance & disclosure. Originally circulated in Chinese on 2026-07-06 (untitled; first line used as headline proxy); drafted with AI assistance under human editorial direction. Translated to English on 2026-08-26 (AI-assisted, human-reviewed). This entry is the thinnest-sourced of Issue 1 — one verifiable anchor (the WEF report) and many [unverified] tags, preserved honestly rather than dressed up. It is translated commentary — not a SigPulse measurement. Our first-party measurements live in the dispatches and the /data/ ledger.