Episode 5: AI Should Remove Noise, Not Judgment
Labor in the Loop
Episode 05 — AI Should Remove Noise, Not Judgment
Content note: This episode is rated NSFKM. Not safe for Karen's mom (or young ears).
About This Episode
Britt just got back from an HR conference where every session she could get to was about AI — and she brought back a lot to unpack. This week she and Karen dig into two ongoing lawsuits reshaping how companies think about AI in hiring: Mobley v. Workday, where a rejected applicant is suing the software vendor instead of the employers who used it, and Harper v. SiriusXM, where a candidate is suing the employer directly. The two trace how bias can sneak into hiring algorithms — through training data, zip codes, employment history, even a single "prototypical" resume used to build a filter — and talk through what's actually fair game to automate versus what companies are handing off to AI that they shouldn't be. Plus: a genuinely great piece of news from Britt, a story about an entrepreneur who ran her business under a male name for decades, and the usual tangents.
What We Cover
[0:00] Cold Open & Conference Recap Britt catches Karen up on the HR conference she attended (funded by the fractional HR client she's been working with) and why she made a beeline for every AI-related session on the schedule.
[4:15] Case One: Mobley v. Workday Derek Mobley applied to over 100 jobs through Workday's platform and was rejected every time — sometimes within minutes. Rather than suing the individual employers, he's suing Workday itself, arguing the software's screening algorithm discriminated based on his race, age, and disability. Britt and Karen break down what Workday actually is (HRIS + ATS), why going after the vendor instead of the employer is such a novel legal strategy, and where the case currently stands (the window to join the class action closed March 7, but the suit is ongoing).
[19:41] Where Bias Actually Sneaks In Karen walks through the mechanics — how a vague preference like "good university" can quietly exclude graduates of historically Black colleges and universities if that's how the training data skews, and why zip code and employment history function as proxies for race even when they look neutral on paper. Includes a detour into redlining, restrictive covenants, and Mapping Inequality, a project digitizing historic redlining maps by city.
[33:50] Case Two: Harper v. SiriusXM A different flavor of the same problem — Arshawn Harper, an IT professional, applied to SiriusXM roughly 150 times and was rejected from all but one, and is suing the employer directly rather than the vendor. Britt and Karen compare the two legal strategies and get into the "not my fault, call the other department" dynamic between vendors and employers when nobody wants to own the liability.
[51:14] Where the Bias in Sirius's Case May Have Come From Britt shares a detail from her conference notes: reportedly, SiriusXM's tool was built using a single resume as its base template — a strong reminder of how easily one "prototypical candidate" can bake bias into a whole system. The two make the case for building screening criteria off the actual job description instead.
[55:58] The Line: Removing Noise vs. Removing Judgment The line from the conference that stuck with Britt: AI should be used to remove noise and repetition, not judgment. She and Karen talk through what that looks like in practice — including a parallel conversation about AI in nonprofit fundraising and donor relationships.
[1:03:10] Britt's News: HR Business Partner Offer Britt shares that she's accepted an HR Business Partner offer at a healthcare organization spanning home health, hospice, and assisted living — a genuine step up from generalist/talent acquisition work, and a role she held out for after months of fractional work gave her the room to be selective.
[1:10:51] The Steve Shirley Story Karen recaps a story from the podcast Cautionary Tales (hosted by Tim Harford) about an entrepreneur who ran her company under the name "Steve" for decades to be taken seriously in business — plus a related story of her own about being advised to publish under a male pseudonym.
[1:21:33] Wrap-Up Quick close-out, plus what everyone's drinking.
Cases Referenced
Mobley v. Workday, Inc. — ongoing; class action window closed March 7
Harper v. SiriusXM — ongoing
Resources Mentioned
Mapping Inequality — digitized historic redlining maps by city
Cautionary Tales podcast, hosted by Tim Harford — source of the Steve Shirley story
Reference to the Enron/Arthur Andersen collapse as the origin of third-party auditing norms
Disclaimer
Neither host is a lawyer. This episode is for entertainment/informational purposes only and isn't legal advice — talk to an actual attorney before making decisions based on anything discussed here.
About Your Hosts
Brittany (Britt) is a certified HR professional based in Boise, Idaho with four years of experience spanning HR generalist work, talent acquisition, and fractional HR consulting through her own practice, Lumi HR LLC. She got into HR the long way, through customer service, restaurants, sales, administration, and a lot of figuring it out as she went. That path is exactly why she cares so much about the people side of work.
Karen Boyd is a values and technology researcher and director of research at a think tank in Southern California. She holds an MBA from the University of California, San Diego and a PhD from the University of Maryland's School of Information, where she studied AI ethics and the future of work. She is, by her own admission, better at studying people than working with them.
Connect With Us
Britt on LinkedIn: https://www.linkedin.com/in/brittanyhumphreys
Karen on LinkedIn: https://www.linkedin.com/in/drkarenboyd/
Lumi HR LLC:https://www.instagram.com/lumihrllc/
Karen’s author account: https://www.instagram.com/karenboydauthor/
Leave Us a Review
If this episode resonated with you, we'd love to hear from you. Leave a review on Spotify or Apple Podcasts. We would love your feedback as we build the show! Please reach out with comments or questions at info@laborintheloop.com.
Labor in the Loop is part career strategy, part tech deep dive, part therapy session between two friends who study work for a living.
