Written by Tim Green, narrated by AI. Listen to the full episode here.
🎙️ Season 1, Episode 22 | Duration: 20:27
Amanda Bowler submitted 50 applications over six months and received zero face-to-face interviews. She is a 48-year-old psychology graduate carrying roughly A$44,000 in HECS debt, and on the available evidence, not a single human being read a word she wrote. Her CV was assessed by software. Her applications were read by nothing.
This episode uses AI voice narration from ElevenLabs Studio.
A Closed Loop With No Author
The number that should stop us is not the volume of applications or the length of the wait. It is the absence of a human decision-maker. Roughly four in five Australian companies now use AI in recruitment at moderate or higher levels, while jobseekers deploy generative agents that scrape boards, rewrite CVs and submit before the applicant has woken. A senior HR leader described it plainly: the bots are botting each other.
The Asymmetry That Matters
The power imbalance is structural. Employer systems render consequential verdicts without accountable authors. Jobseeker systems produce tailored applications at near-zero cost. Both are automated. Only one side decides who gets rejected and who does not.
Bias Without Intent
The Mobley v Workday class action, filed in February 2023 by Derek Mobley, an African American man over 40 with depression and anxiety, alleges he was rejected from more than 100 positions through Workday's screening software, sometimes outside business hours and within an hour of applying. Around a billion applications were rejected through the software during the relevant opt-in period. Roughly 14,000 people opted in to the class.
Evidence of Disparate Impact
Research on large language model preferences shows consistent patterns: race-linked and gender-linked resume model preferences that replicate existing discrimination without any intentional design. Scaled AI screening interviews add another layer, rendering decisions that are simultaneously consequential and opaque.
What a Right to a Human Decision Would Require
The thirty-second rejection is not painful because it is fast. It is painful because its speed discloses that nothing was considered. If a bank refuses a loan, there is a decision-maker, a reason and a complaints process. If an AI screening tool rejects an application, there is none of these things.
Proposed Remedies
Point-of-application disclosure that a person is interacting with an AI system. Reversible machine rejections by a named individual. Audited human sampling of rejected applications. Published reasons and age-band rejection rates. Portable identity verification to restore the signalling function that credentials once provided. What applicants are owed, above all, is a reader.
Key Sources
- Australian Responsible AI Index 2025 - Fifth Quadrant / National AI Centre
- Mobley v Workday, Inc. - US District Court, Northern District of California
- SEEK Employment Report, July 2026 - SEEK
- EU AI Act Regulation 2024/1689 - EUR-Lex
Listen to the Full Episode
🎧 Rejected in 30 Seconds: When No Human Reads Your Application | Duration: 20:27
Subscribe on Apple Podcasts, Spotify, or your favourite app.
SmarterArticles is written by Tim Green, narrated by AI via ElevenLabs Studio. New episodes every Monday. Follow @humanin_theloop for updates.
Top comments (0)