Most companies believe they fixed bias in hiring when they added AI to the process. They didn’t. They scaled it.
I sat down with Shauna Petterson — a legal technology recruiter and For Humanity Fellow — on The Human Code podcast to unpack what’s actually happening inside AI-powered hiring systems. Her perspective is both deeply researched and practically grounded, and what she shared challenges some of the most widely held assumptions in the industry.
Here’s what came out of that conversation.
The Promise Was Never Matched by the Reality
AI hiring tools were sold as the neutral, scalable solution to a deeply human problem. Remove unconscious bias. Streamline candidate review. Surface the best people faster.
The promise made sense. The execution didn’t.
Those systems were built on historical hiring data — years, sometimes decades, of human decisions, complete with their embedded prejudices, blind spots, and cultural assumptions. The AI didn’t arrive neutral. It arrived pre-loaded with the same flawed patterns it was supposed to fix.
And because automation works at scale, those patterns now play out invisibly and at unprecedented speed. Bias that once affected dozens of decisions a day now affects thousands.
The Dead Internet Theory Has Come for Recruiting
There’s a concept called the Dead Internet Theory — the idea that a growing portion of online interaction isn’t humans talking to humans anymore. It’s bots responding to bots, algorithms optimizing against other algorithms.
Hiring has its own version of this.
The modern recruiting ecosystem is saturated with ghost jobs — positions posted but never intended to be filled — bot-generated applications flooding applicant tracking systems, and resume-optimization tools designed to game keyword filters rather than communicate actual qualifications.
The result: a market full of synthetic activity where legitimate opportunities get buried, qualified candidates get filtered out before a human ever sees their application, and both sides are optimizing for the machine rather than for each other.
Who Gets Left Behind: A New Kind of Divide
The gamification of hiring hasn’t landed equally. There’s a growing socioeconomic divide between candidates who can access the tools to compete and those who can’t.
The Resource Gap
On one side: candidates with time, money, and technical knowledge to build custom application tools, subscribe to premium optimization services, and invest hours gaming each individual ATS system.
On the other: candidates who lack those resources and are therefore filtered out before a human ever evaluates them.
This isn’t a talent gap. It’s an access gap that looks like a talent gap.
The Neurodivergent and Disability Penalty
Video analysis tools that assess micro-expressions, speech cadence, and eye contact patterns systematically disadvantage neurodivergent and disabled candidates — people who may be exceptionally qualified but whose communication style doesn’t match the algorithmic model of professional behavior.
The irony: these tools claim to reduce bias while introducing entirely new categories of it.
The Gamification Problem — and What It Costs Organizations
When hiring systems select for the ability to navigate ATS filters rather than the ability to do the actual job, organizations lose something they can’t measure on a scorecard: real capability.
The candidates who rise to the top are often those most skilled at appearing qualified within a particular system. The candidates who get filtered out are sometimes those who would actually perform best.
This is a structural problem, not a personnel problem. And it compounds: bad hires lead to worse hiring data, which leads to worse AI recommendations.
Your Network Is Still Your Net Worth
Despite all of it, the data on how people actually land jobs hasn’t changed: the overwhelming majority find work through personal relationships, not algorithms or job boards.
Technology didn’t replace the power of human networks. It displaced attention from them — often toward activities with lower return.
Job seekers who invest in building genuine professional connections consistently outperform those who spend equivalent time perfecting keyword density. This remains true even as AI tools have multiplied.
Building Better Tools: Why Oversight Matters
One of the structural issues Shauna identified is organizational: the cultures that deploy flawed hiring tools often suppress honest feedback about them. When a system is a VP’s pet project or a significant budget line item, the incentive to surface its failures is low.
Meaningful improvement requires something most organizations don’t currently have: oversight committees with genuine decision-making authority, meaningful diversity, and actual power to challenge existing tools.
It also requires acknowledging a difficult truth. Bias can’t be eliminated from AI hiring systems. It can only be actively managed and contained. Tools that claim otherwise are either uninformed or misleading — and organizations that believe those claims are setting themselves up for the exact failures the tools were supposed to prevent.
What Job Seekers Can Do Right Now
Shauna’s practical advice is grounded in what’s actually been shown to work:
Prioritize relationships over tools. Genuine human connection still outperforms any AI-optimized application strategy. Invest time building real professional networks before investing in software subscriptions.
Focus on continuous learning. Developing real skills matters more than ATS optimization. The goal is to be worth hiring — not just good at looking like it.
Acknowledge the psychological cost. Job searching is difficult under normal circumstances. Navigating AI-mediated rejection at scale makes it significantly harder. Taking that seriously — and finding support — isn’t weakness. It’s an accurate response to a real challenge.
And if AI tools feel overwhelming: start with human-centered approaches first. Exhaust those before turning to algorithmic workarounds.
The Bigger Picture
The gap between what AI hiring technology promised and what it delivers isn’t randomly distributed. It falls disproportionately on the populations that already face the steepest barriers to opportunity: people from lower socioeconomic backgrounds, neurodivergent candidates, and people of color.
Closing that gap isn’t primarily a technical challenge. It’s an accountability challenge. It requires asking who sits in audit and oversight roles, who has the authority to challenge systems that aren’t working, and who bears the cost when they fail.
Get Involved: For Humanity
Shauna’s organization, For Humanity (forhumanity.app), develops ethical AI audit frameworks and makes them openly available — free coursework, weekly working sessions on AI ethics and hiring practices — for anyone who wants to engage seriously with these questions.
Whether you’re a hiring manager, an HR technology buyer, or a job seeker navigating this landscape, it’s worth exploring.
This Conversation Lives on The Human Code
This piece is drawn from an episode of The Human Code, the FINdustries podcast hosted by Don Finley. Every episode explores the intersection of technology, ethics, and human experience — with practitioners working through these questions in real organizations.
Subscribe on YouTube, Spotify, Apple Podcasts, or wherever you listen.
Work With FINdustries
FINdustries helps organizations think clearly about AI — where it creates value, where it introduces risk, and how to build governance structures that make it trustworthy over time.
If you’re an executive reconsidering your hiring technology stack, or a leader who wants to deploy AI in ways your organization can actually stand behind, let’s talk.
Visit findustries.co to learn more.
