Social Listening Experiment

AI Changed the Signal in Hiring.
Here's What People Are Saying.

604
Comments Analyzed
45%
Expressed Frustration
500+
Avg. Apps Mentioned
The Core Problem

"Someone who uses AI now looks just as hireable as someone who doesn't. The signal employers relied on is gone."

Research: "Making Talk Cheap: Generative AI and Labor Market Signaling" by Anaïs Galdin (Dartmouth Tuck) & Jesse Silbert (Princeton) — Read Paper

What This Is

I made a video breaking down research on how AI broke hiring. It hit 320K views. I wanted to understand what resonated, so I scraped all 604 comments and analyzed them with Claude Code. Here's what I found.

Sentiment Analysis

The overwhelming sentiment? Frustration and validation. People aren't just complaining—they're relieved to finally have data confirming what they've been experiencing.

45%
Frustrated / Angry
30%
Validated / Relieved
15%
Cynical / Resigned
10%
Hopeful
Sentiment Distribution
Category What It Sounds Like
Frustrated / Angry Direct criticism, personal grievances, system failures
Validated / Relieved "Finally someone said it" — externalized blame for struggles
Cynical / Resigned Gallows humor, fatalistic acceptance, "it's all broken"
Hopeful Seeking solutions, constructive suggestions, optimism

What People Are Actually Talking About

Five dominant themes emerged from the comment data. The most prevalent: raw, personal stories of job search trauma.

Job Search Trauma
35%
AI vs. AI Hypocrisy
25%
Death of "Signal"
18%
Gaming the System
12%
Ghost Jobs / Distrust
10%

Job Search Trauma (35%)

Personal narratives of extended, unsuccessful job searches. Commenters frequently cited specific metrics: hundreds of applications, months or years of searching, minimal human contact throughout the process.

AI vs. AI Hypocrisy (25%)

The perceived double standard: companies deploying AI to screen candidates while criticizing applicants for using AI to prepare materials. Frequently described as an "arms race" with no winner.

Death of "Signal" (18%)

Direct engagement with the research finding: AI has eliminated the traditional signals (effort, customization) that differentiated candidates. Everything looks the same now.

Gaming the System (12%)

Shared workarounds: hidden text prompts for AI scanners, altering demographic markers, strategic keyword stuffing. Adaptive behaviors in response to perceived algorithmic gatekeeping.

Ghost Jobs & Distrust (10%)

Recurring references to "ghost jobs" (posted positions with no intention to hire) and broader distrust of corporate hiring practices. Several commenters advocated for regulation.

In Their Own Words

Selected for clarity and representativeness of broader themes. All quotes are verbatim.

"Companies are complaining they can't hire because applicants are terrible. But you're getting what your AI is giving you!"
"You have no idea how much this video means to me after 500 job applications and 1 interview. I spent months thinking it was just me."
"They set hiring standards so high only liars meet them. Hence companies only get liars in the door."
"If they don't want candidates using AI, then don't use it yourselves."
"The lack of training up is indicative of crappy leadership. No one upskills anymore. They pigeon hole people and then wonder why they didn't grow and seek from outside. Just dumb."
"Ghost jobs should be illegal."

Who Showed Up

Based on self-disclosed information, commenters fell into several overlapping groups—different lenses through which people engaged with the content.

By Hiring Position

  • Active job seekers — the majority; sharing application counts, timeline frustrations
  • Recently employed — reflecting on their own difficult searches
  • Recruiters & HR — small but notable; some defensive, some agreeing
  • Hiring managers — rare; a few admitting the system is broken

By Career Stage

  • Recent graduates — entry barriers, "need experience to get experience"
  • Mid-career (5-15 yrs) — overqualification filtering, being "too senior"
  • Senior professionals — ageism concerns, skills discounted

By Industry

  • Tech / Software — most vocal; senior devs unable to get callbacks
  • Corporate / Admin — project coordinators, operations roles
  • Former DEI pros — field collapsed; forced into unrelated work
  • Creative / Marketing — AI concerns compounding job scarcity

Methodology

  1. Data Collection

    604 comments extracted from the TikTok video using a comments scraper. Data exported as CSV on January 12, 2026, capturing comment text, timestamps, usernames, and engagement metrics.

  2. AI-Assisted Analysis

    Comments analyzed using Claude (Anthropic) to identify patterns, themes, and sentiment. The AI processed all comments, extracting key phrases, categorizing tone, and flagging recurring topics.

  3. Sentiment Classification

    Four sentiment categories: Frustrated/Angry, Validated/Relieved, Cynical/Resigned, and Hopeful. Percentages are approximate estimates based on pattern distribution.

  4. Thematic Coding

    Five major themes identified through qualitative analysis. Individual comments could address multiple themes.

  5. Quote Selection

    Representative quotes selected based on engagement, clarity, and perspective diversity. All verbatim.

Limitations: This is qualitative analysis, not a rigorous quantitative study. Percentages are estimates. Self-selection bias is inherent in social media data. And yes—the irony of using AI to analyze comments about AI's impact on hiring is acknowledged.