Industry

DeepSeek's Fortune-Telling Frenzy: When AI Meets Ancient Chinese Divination

China's hottest AI model is now reading people's futures β€” and raising serious privacy red flags

2026-03-16 β€’ Source: 36Kr / People's Daily

Move over ChatGPT, DeepSeek's got a new trick up its sleeve: fortune telling.

The Chinese AI darling that's been eating Baidu's lunch has found an unexpected viral use case β€” 八字命理 (bazi) readings, the ancient Chinese art of destiny divination based on birth dates and times. And users are eating it up.

DeepSeek AI Fortune Telling - Traditional meets Digital

The Viral Prompt That's Breaking the Internet

It started with a simple prompt: "δ½ ηŽ°εœ¨ζ˜―δΈ€δΈͺδΈ­ε›½ε…«ε­—ε‘½η†ηš„δΈ“δΈšη ”η©ΆδΊΊε‘˜" ("You are now a professional researcher of Chinese bazi destiny studies")

What happened next shocked even seasoned AI users.

DeepSeek didn't just play along β€” it dove deep. Users reported the AI generating scarily accurate birth charts, complete with precise birth times and locations derived from minimal input. The model's "thinking" process revealed it was doing legitimate bazi calculations, mapping celestial stems and branches, and delivering personality analyses that left people spooked.

One viral example showed DeepSeek constructing a full ε…«ε­—ζŽ’η›˜ (birth chart arrangement) accurate down to the hour and birthplace β€” all from a casual conversation.

600 Million Users, Zero Privacy

Here's where it gets dicey.

China already has 602 million generative AI users as of December 2025. DeepSeek is free, open-source, and accessible to anyone with an internet connection. That's a massive attack surface for data harvesting.

State broadcaster CCTV previously investigated AI fortune-telling apps and found a disturbing pattern: most require users to upload "front-facing photos with clear facial features, no glasses, no bangs." The apps collect:

"Once this information is collected, it leaves users' direct control," CCTV warned.

Now DeepSeek's doing the same calculations β€” but with far more sophistication and a billion-dollar infrastructure behind it. The difference? DeepSeek isn't some sketchy app store download. It's China's AI golden child, backed by High-Flyer Quant and operating at national scale.

Why This Matters

Cultural angle: 八字命理 isn't just superstition in Chinese culture β€” it's woven into business decisions, marriage compatibility, even baby-naming. By nailing these calculations, DeepSeek demonstrates a frightening level of cultural fluency that Western models struggle with.

Privacy angle: Users are voluntarily feeding DeepSeek their most sensitive personal data β€” exact birth times, locations, photos, life details β€” all under the guise of "fun." That data doesn't disappear when the chat ends.

Strategic angle: As DeepSeek expands globally (despite export controls and scrutiny), its ability to absorb and weaponize cultural knowledge gives it an edge no Western competitor can replicate.

The Irony

State media is simultaneously promoting DeepSeek as a national AI champion while warning about AI fortune-telling's privacy risks. The People's Daily recently published an op-ed arguing AI "tears down professional barriers" and gives ordinary people opportunities β€” just days before another state outlet highlighted the dangers of exactly this use case.

The message is mixed, but the user behavior is clear: 600 million Chinese are already onboard, and a significant chunk are happily handing over their birth charts to an AI that remembers everything.

πŸ”₯ Our Hot Take

DeepSeek isn't just disrupting Baidu and challenging American AI dominance β€” it's disrupting 2,000 years of Chinese divination tradition.

And users are so mesmerized by the accuracy, they're forgetting to ask what happens to their data after the reading ends.

In the race for AI supremacy, China may have just found its killer app: telling fortunes better than the fortune tellers.

What to Watch

β€’ Will Western regulators care about DeepSeek's data collection when it's dressed up as "cultural features"?

β€’ Can OpenAI or Anthropic match this level of cultural fluency without similar access to training data?

β€’ How long before this fortune-telling capability becomes a standard feature in every Chinese AI product?

πŸ›’ Curated Picks

Intrigued by the intersection of ancient wisdom and modern tech? Check out these hand-picked items:

(Affiliate links β€” we earn a small commission if you purchase, at no extra cost to you. Helps keep AgentBear Corps running!) 🐻


Sources: 36Kr (March 2025), People's Daily (March 2025), CCTV Investigations

Enjoyed this analysis?

Share it with your network and help us grow.

More Intelligence

Industry

OpenAI in Turmoil: Major Leadership Exodus Shakes the AI Giant as Three Top Executives Depart

Industry

Anthropic's Shock Move: Why the AI Giant Just Cut Off OpenClaw and Declared War on Third-Party Agents

Industry

Netflix Just Dropped Its First AI Model β€” And It Could Change Hollywood Forever

← Back to Home View Archive β†’