What Now? With Trevor NoahHilke Schellmann: Is the Algorithm Hiring the Wrong People?
At a glance
WHAT IT’S REALLY ABOUT
AI hiring tools scale bias, reject talent, and surveil workers.
- Schellmann argues that traditional hiring (resumes and interviews) is already weak at predicting performance, and AI often scales those same flawed assumptions rather than fixing them.
- Many large employers use AI early in the funnel to screen resumes, rank candidates, and run one-way video interviews or gamified assessments—often without proving these tools actually predict job success.
- Real-world cases show algorithmic proxies can penalize protected groups indirectly (e.g., Amazon downgrading resumes containing “women”), creating discrimination that’s harder to detect and litigate.
- Applicants increasingly respond with their own AI (auto-applications, AI-written resumes), turning hiring into “AI vs. AI” and raising questions about the purpose and legitimacy of the process.
- Beyond hiring, AI-driven workplace surveillance and productivity scoring can pressure employees into “productivity theater,” enable intrusive monitoring, and potentially justify firing decisions with plausible deniability.
IDEAS WORTH REMEMBERING
5 ideasAI didn’t break hiring—hiring was already broken.
Schellmann frames AI as exposing how poor interviews and resumes are at measuring real capability, often rewarding confidence and familiarity rather than competence.
Automated screening commonly rejects qualified candidates at scale.
She cites survey results where nearly 90% of leaders at companies using AI acknowledged their tools reject qualified applicants, yet adoption continues for efficiency reasons.
Many AI hiring signals are unjustified proxies, not job-relevant evidence.
Examples include keyword-driven boosts/penalties (e.g., “women,” “softball,” certain names/places) and assessments like rapid space-bar tapping that claim to infer personality or risk tolerance.
Bias can become more “defensible” when it’s automated.
Using vendors and opaque models can create plausible deniability, making it difficult for applicants or workers to prove discrimination even when outcomes are systematically skewed.
AI hiring is increasingly an arms race between applicants and employers.
With candidates using LLMs to tailor resumes and even auto-apply, employers respond with more automated filters—risking a process optimized for gaming rather than truth-finding.
WORDS WORTH SAVING
5 quotesSo we have this, like, competence versus confidence problem.
— Hilke Schellmann
Like, you know, it's kind of like- when I'm in a job interview and, and I smile, and people... You know, a facial, uh, emotion-scanning algorithm would say, like, "Oh, yeah, she's totally happy. She's smiling." And I'm like, "I'm fucking nervous."
— Hilke Schellmann
If your company uses AI tools, um, do they reject qualita- qualified, do they reject qualified candidates? And almost 90%, um, of the leadership said yes.
— Hilke Schellmann
If you had the word, uh, woman or women... um, on your resume, you got downgraded.
— Hilke Schellmann
Oh my God, how did you know that? I think it's plausible deniability, um, of the companies that use it and buy from the vendor...
— Hilke Schellmann
High quality AI-generated summary created from speaker-labeled transcript.
