- Chen Yu received a message: "Your resume has been submitted to our database." He never applied. An AI recruitment platform had autonomously submitted his resume at 3:42 AM — 97.3% confidence.
The system decided he should leave his job based on: browsing history (three midlife-crisis articles), declining commit frequency (his father was in the hospital), dormant professional network (also his father). Every data point pointed to "he should leave." But the algorithm couldn't see the one thing that mattered: his father had liver cancer.
So Chen wrote a "digital scarecrow" — a Python script that visited twenty recruitment sites, introduced typos in his resume, sent ambiguous messages. Make the signal so noisy no HR would take it seriously.
A month later, his recommendation notifications dropped to zero. The algorithm didn't get smarter. It calculated risk-reward ratio — and a noisy signal was bad business. The algorithm doesn't care about truth. It cares about conversion rate.
A story about the only survival strategy left: making the algorithm decide you're not worth it. Read the full bilingual story on the blog.
This article first appeared on my tech blog
Top comments (0)