Job seekers learn to please the algorithm as AI interviews take over campus recruiting
创业邦 · wechat · 2026-11-19
A New Weekly feature documents how AI has taken over China's campus recruiting — and why applicants hate it:
- Resume screening: ATS keyword filters force applicants to write for the machine — keywords in the top third, quantified results ("boosted efficiency 30%"), then a final AI pass to strip out the "AI flavor."
- The AI interview: mainstream systems (Beisen, Zhilian, Liepin, Nowcoder) share question banks, so "every AI interviewer has the same face." Answers must stretch to 4.5 minutes in STAR format; ASR errors, premature topic cuts and buggy camera checks abound.
- Asymmetric effort: after 40–60 minutes of performative answers, applicants get no feedback and are silently filtered out; some firms even demand videos of applicants running and squatting — read by many as an "obedience test."
- Magic vs. magic: cheating apps whisper real-time answers; eye-fixation plugins keep your gaze pinned to the camera. Recruiters counter that AI can't tell nervousness from cluelessness.
The piece argues AI interviews select for the most "standard" candidate, not the most capable — turning hiring from mutual evaluation into a one-way tribunal.
More from AGI Musings
- Stratechery Interview: Eric Seufert on Why Ads Win Agentic Commerce — Stratechery · 2026-11-19
- Quanta: Is 'Attention' the Essence of Intelligence? Tracing the 2014 Breakthrough — MacrinePhD · 2026-11-19
- Neuroscientist Anil Seth to deliver Tans Lecture making "The Case Against Conscious AI" — anilkseth · 2026-11-19
- How many papers go extinct with model deprecation — and how many are littered with needless classifiers? — repligate · 2026-11-19
- Researcher accuses Anthropic of chasing global AI monopoly while alarmism doesn't match its actions — kevinnbass · 2026-11-19
- Practical AI podcast: human-centered AI adoption in banking with Scotiabank — Practical AI · 2026-11-19