The Scene

Last Wednesday I scrolled my WeChat Moments for 20 minutes, then froze on a travel photo — the caption was so polished it didn't feel real. I couldn't tell if a human wrote it or AI.

What It's Called + Who's Talking About It

There's a new English term for this: "AI-blind" — like color blindness. The more AI content floods our feeds, the worse we get at spotting it. The original observation came from a blogger named Cymerys. My freelance writer friend Ailin came to me last week venting — she spent 6 hours hand-writing a piece for a client, and the client's first response was "Did you use AI for this?" That's where our industry is right now: trust is quietly being discounted, and nobody's noticing.

Cost to Replicate + First Step

This isn't about buying a new tool. It's more like "pause and think for 30 seconds." Here's what I do —

  • Before posting to Moments or Xiaohongshu, add one specific scene: "Today at the 7-11 on XX Road, I saw XX." The more specific, the less it reads as AI.
  • In deliverables to clients, intentionally leave 1-2 casual verbal pauses.
  • Stop chasing perfection — perfection now reads as AI.

Money: 0. Time: 5 min/day. Tech barrier: you already know how to post. First step: tomorrow, post one Moments update with a specific place + specific action, and see who replies.

Advice by Stage

Just starting (no clients yet): Don't stress about this yet. Your job now is to be remembered, not suspected of being AI. Being intentionally rough is actually an advantage — when I started out, I deliberately didn't edit my photos, and a few clients noticed me immediately.

Have 1-2 clients: Add 1-2 sentences of "my own small observation" into your deliverables, so the client actually sees a person. Next time they ask "Did you use AI?", you'll have something real to answer with.

Scaling (team / batch content): This one really matters. I made this exact mistake before — hired a writer whose work all smelled like AI, and client churn jumped 30% in a single month. Now we require every piece to pass a "human smell check": read it out loud, if it doesn't roll off the tongue, rewrite.

This tool/method isn't for everyone. If you're not currently anxious about being accused of using AI, you don't have to try it now — but I'd suggest at least spending 5 minutes a day asking "does what I post sound human enough?"