← Back to stories A vintage camera on brown paper with 'SOCIAL MEDIA MARKETING' text, perfect for marketing visuals.
Photo by Eva Bronzini on Pexels
凤凰科技 2026-03-29

It's time to invite Jiang Shuying (江疏影) to be a spokesperson

A celebrity could make algorithmic mechanics public

China’s short‑video behemoth Douyin (抖音), operated by ByteDance (字节跳动), is often described as addictive—but what if the public could be guided through exactly why? It has been reported that an ifeng analysis dismantles the mechanics behind that “can’t stop” feeling, arguing that a famous, trusted face like actress Jiang Shuying (江疏影) could help translate technical nuance into public literacy. Who better to bridge entertainment and public debate than a popular cultural figure who speaks to mainstream audiences?

AI, the Hooked model, and the anatomy of “just one more”

The story unpacks how Douyin’s feed is not random noise but an AI‑optimized implementation of the Hooked model popularized by Nir Eyal: trigger, action, variable reward, investment. Reportedly, the platform uses a content matrix and early‑session tests to build an accurate initial profile within the first few minutes; the platform watches micro‑behaviours—three‑second watch thresholds, swipe speed, micro‑pauses—and tunes “variable rewards” so users stay in the sweet spot of “a little satisfied, not satisfied enough.” The result is not mere randomness but emotion‑aware sequencing that raises surprise rates without delivering uniformly perfect content.

Why this matters to product builders and regulators

For Western readers: this is part product design, part behavioral economics, and part machine learning. It also sits inside an active regulatory context—China has tightened rules on recommendation algorithms and youth protection, and platforms worldwide face political scrutiny over engagement‑driven design. For product managers the lessons are practical: design the cold‑start experience, manage variable expectations, and compile a “friction list” to remove tiny decision points. For the public, the lesson is civic: informed debate about algorithmic harms needs accessible interpreters, not just academics and engineers.

A small proposal with outsized impact

Inviting a mainstream figure like Jiang Shuying to front a campaign or public service conversation would not solve the technical problems, but it could shift the discourse—turn opaque mechanisms into everyday terms and prod both companies and regulators toward clearer guardrails. The ifeng piece promises a four‑step toolkit for analysing addictive dynamics; perhaps the most potent next step is simple: educate at scale.

AI
View original source →