Class Action Filed in Sixth Circuit Against Sirius XM Radio

When a lawsuit mixes artificial intelligence, hiring, race discrimination, and a major media brand, it is bound to make noise. That is exactly what happened with Harper v. Sirius XM Radio, LLC, a federal case filed in Michigan that has put SiriusXM under a fresh legal spotlight. To be precise, the complaint was filed in the U.S. District Court for the Eastern District of Michigan, which sits within the Sixth Circuit. So the headline shorthand works, but the legal geography deserves a quick tune-up: this is a district-court case inside the Sixth Circuit, not an appeal already pending before the Sixth Circuit Court of Appeals.

The lawsuit matters because it is not just another employment dispute with a few spicy allegations and a stack of pleadings tall enough to flatten a desk plant. It speaks to a much larger question hanging over modern hiring: when employers use software to screen people, rank résumés, and narrow candidate pools, who carries the legal risk if the machine behaves badly? According to the complaint, SiriusXM’s hiring process may have done exactly that by using technology in a way that allegedly disadvantaged Black applicants. SiriusXM, for its part, has contested the case and moved early to challenge it.

For employers, compliance teams, HR leaders, recruiters, and job seekers, this case is worth watching. It sits at the crossroads of Title VII, algorithmic decision-making, vendor accountability, adverse impact analysis, and the basic human desire to not be rejected by what feels like a robot wearing a business-casual disguise.

What the SiriusXM lawsuit actually says

The complaint was filed by Arshon Harper against Sirius XM Radio, LLC. At its core, the case alleges that SiriusXM used an AI-powered hiring system in a discriminatory way. Reports discussing the complaint say Harper alleged that SiriusXM relied on the iCIMS Applicant Tracking System and that the technology screened, scored, or downgraded applications in a manner that disproportionately harmed African American applicants.

That claim is important for two reasons. First, the lawsuit is framed as a proposed class action, meaning it seeks to go beyond one applicant’s experience and represent a broader group of similarly situated people. Second, the theory is not limited to obvious, old-school discrimination. It focuses on algorithmic discrimination, which is often more subtle and harder to spot. Instead of a hiring manager typing something reckless into an email, the concern is that software may use inputs or patterns that function as proxies for race, such as zip code, educational background, employment history, or other data points that can reflect historical inequality.

According to commentary on the filing, Harper claimed he applied for roughly 150 roles at SiriusXM, was rejected repeatedly, and received only one interview before being turned down again. That alleged pattern is central to the lawsuit’s theory. The complaint argues that the hiring system did not simply help organize applications, but effectively shaped who received real consideration and who never got past the digital velvet rope.

Why the “Sixth Circuit” label matters

Headlines love shortcuts. Lawyers do not. The distinction matters because a case filed in a district court is still in its fact-building stage, where pleadings, motions, discovery fights, and procedural battles usually happen. A case in the court of appeals is at a very different stage, where a trial-court record already exists and the fight is mostly about legal error.

Here, the lawsuit was filed in the Eastern District of Michigan, which falls under the Sixth Circuit umbrella. That means any later appeal would likely go to the U.S. Court of Appeals for the Sixth Circuit. It does not mean the Sixth Circuit has already ruled on the merits. That distinction is not just technical trivia for legal nerds and extremely enthusiastic footnote readers. It affects how readers should interpret the case. Right now, these are allegations in active litigation, not findings of fact.

As of early January 2026, SiriusXM had answered the complaint and filed a motion for judgment on the pleadings, signaling that the company is trying to knock out or narrow the case before full-blown discovery takes over everyone’s calendar.

The legal theory behind the class action

Disparate treatment and disparate impact

Employment discrimination law has more than one lane, and this case appears to travel in at least two of them. One lane is disparate treatment, which is the classic theory that an employer intentionally treated someone worse because of a protected trait such as race. The other lane is disparate impact, which focuses on a seemingly neutral practice that disproportionately harms a protected group and cannot be justified as job-related and consistent with business necessity.

That second theory is where AI hiring litigation gets especially interesting. A software tool may look neutral on paper. It may not ask for race. It may not display race. It may not even “know” race in any direct sense. But if the system relies on variables or training patterns that mirror older bias, the result can still be discriminatory. In plain English, the algorithm does not get a legal free pass just because it never rolled its digital eyes.

The vendor question

The case also raises the recurring question of how much blame belongs to the employer and how much belongs to the software vendor. Federal guidance strongly suggests employers cannot shrug and point at the vendor like a sitcom character caught next to a broken printer. If an employer uses automated tools to make or inform hiring decisions, it can still face liability if those tools have an unlawful discriminatory effect.

That principle has become even more important as ATS platforms and AI recruiting products market themselves as faster, smarter, more efficient, and more “equitable.” The promise is seductive: less manual screening, quicker matching, more scalable recruiting. The legal risk is less glamorous: if the tool filters out the wrong people for the wrong reasons, the employer may still own the problem.

Why this case fits a bigger national trend

The SiriusXM lawsuit did not appear out of nowhere. It arrived during a period of escalating scrutiny over AI in employment. The EEOC has repeatedly said that federal anti-discrimination laws apply when employers use AI or other automated technologies in recruiting, screening, hiring, promotion, pay, and termination decisions. The agency has also emphasized that even a facially neutral tool can violate the law if it creates an unjustifiable disparate impact.

That guidance matters because companies often treat hiring software as administrative plumbing. But regulators do not see it that way. If the tool influences who gets seen, who gets scored, or who disappears into the résumé void, it becomes part of the selection procedure. And selection procedures are not exempt from Title VII just because they arrive with dashboards, analytics, and a cheerful SaaS sales pitch.

Another reason the SiriusXM case feels significant is the broader litigation wave involving AI hiring tools. The closely watched Workday litigation has already helped frame the debate over whether software vendors themselves can be pulled into anti-discrimination claims. Even when the theories differ, the direction of travel is obvious: courts, agencies, and plaintiffs’ lawyers are no longer treating algorithmic hiring as futuristic science fiction. It is regular litigation now, just with more data fields.

Where iCIMS and similar platforms enter the picture

iCIMS describes itself as an enterprise-grade applicant tracking system and AI-powered recruitment platform. Its public marketing highlights tools designed to streamline recruiting, embed AI into hiring workflows, improve candidate matching, and support faster decision-making. It also says its AI tools aim to support equitable and transparent hiring decisions.

That type of positioning is common across the HR-tech market. Vendors want to sell efficiency without selling legal panic, so the messaging usually blends speed, fairness, and responsible AI language into one neat package. The SiriusXM lawsuit illustrates why those claims now face real-world pressure tests. If a system is marketed as sophisticated enough to help make better hiring decisions, plaintiffs will argue it is sophisticated enough to be examined when those decisions appear skewed.

None of that proves the allegations in Harper. But it does explain why the case matters beyond one company. It could force deeper questions about how employers validate tools, what data they use, whether adverse-impact testing happens before launch or only after trouble begins, and how transparent both employers and vendors are when candidates challenge outcomes.

What SiriusXM may argue in response

SiriusXM has not been found liable in this case, and any fair analysis has to leave room for the company’s defenses. In a dispute like this, defendants often argue that the complaint relies on speculation, not hard statistical proof. They may say the software did not make final decisions, that human review remained part of the process, or that rejected applicants were not similarly situated in the way the complaint suggests.

The company may also push back on the class-action structure itself. Class actions in employment cases are not automatic. Plaintiffs usually must show enough common questions and shared circumstances to justify treating a broad group as one class. If hiring decisions varied by role, manager, location, qualifications, or workflow, SiriusXM could argue that the claims are too individualized for class treatment.

And then there is the procedural front. A motion for judgment on the pleadings is an early-stage attack. It generally says, in effect, “Even taking the complaint seriously, this case still should not proceed as pleaded.” That does not decide the facts, but it can shape how much of the lawsuit survives long enough to reach discovery.

Why employers should care even if they never used SiriusXM or iCIMS

This case is really about governance. Employers using AI in hiring should assume regulators and plaintiffs will ask some basic questions: What exactly does the tool do? What data does it rely on? Was it validated for the jobs at issue? Has anyone tested for adverse impact? Were less discriminatory alternatives considered? Can the employer explain the results in human terms, or is the response essentially, “Well, the algorithm vibes were strong”?

That is where federal guidance and local rules start to converge. New York City’s Local Law 144, for example, requires bias audits and disclosures for certain automated employment decision tools. Federal agencies have also made clear there is no magical AI exception to civil rights and consumer protection laws. If anything, the message from Washington and major local regulators is refreshingly blunt: high-tech tools still have to obey old-fashioned law.

For compliance teams, the practical lesson is simple. Audit before deployment, not after a complaint lands. Document the business necessity. Review vendor contracts carefully. Demand clarity about how recommendations are generated. Monitor outcomes by protected class where legally appropriate. And do not let anyone in the room say, “The software handles that,” as if the software also passed the bar exam.

SiriusXM’s broader litigation backdrop

The Harper case is separate from SiriusXM’s other legal issues, but it does not arise in a vacuum. SiriusXM has disclosed in securities filings that it faces class actions and mass arbitrations tied to pricing, billing, and cancellation practices. It also disclosed a 2025 agreement to settle a do-not-call class action, with $28 million paid into a settlement fund in 2026. In another unrelated matter, a New York judge found the company liable in a case brought by the state attorney general over subscription-cancellation practices.

None of those separate matters proves anything about the AI hiring allegations in Harper. Different facts, different laws, different claims. Still, they reinforce an important point: large subscription-driven companies live under constant scrutiny when their systems, scripts, or workflows affect consumers or applicants at scale. Once process becomes product, process also becomes litigation risk.

The experience side of the story: what this topic feels like in real life

Whether Harper ultimately wins or loses, the lawsuit taps into a very familiar modern experience: talented people sending applications into an online system and hearing nothing back except an automated rejection that arrives with the warmth of a parking ticket. Many applicants already suspect that software is making meaningful decisions long before a human recruiter ever sees their résumé. The SiriusXM case gives legal vocabulary to that frustration.

From the applicant’s perspective, the experience can feel almost surreal. You tailor your résumé, match the keywords, polish your work history, answer the screening questions, and hit submit. Then, sometimes within hours, you get a rejection so fast it feels less like evaluation and more like a trapdoor. After enough rounds of that, people begin to wonder whether they were rejected because they were unqualified, because the job was never truly open, because the system wanted a different profile, or because the software was drawing invisible conclusions from background details that should never have been outcome-determinative in the first place.

Recruiters and HR teams face a different version of the same pressure. They are often buried under application volume, asked to move faster, and expected to show cleaner metrics with fewer staff hours. AI tools promise relief. They can sort applicants, suggest matches, rank résumés, automate communication, and make overwhelmed teams feel like they finally found a life raft. But a life raft is not much help if it quietly leaks compliance risk into every hiring cycle.

Hiring managers are also learning that convenience comes with tradeoffs. A black-box recommendation engine may seem helpful when it surfaces “top candidates,” but it can be difficult to challenge or interpret when someone asks why a qualified person was screened out. That is the real tension in this entire area. Employers want efficiency. Candidates want fairness. Regulators want accountability. Vendors want to sell innovation. Courts, meanwhile, get to sort through the awkward aftermath when those goals collide.

There is also a trust problem brewing. Job seekers increasingly assume they are being judged by systems they cannot see and criteria they cannot test. Employers, on the other hand, often assume vendors have already addressed bias, explainability, and validation. Sometimes both sides are relying on faith when what they really need is evidence. Lawsuits like the one against SiriusXM push that gap into public view.

That is why this case resonates far beyond one plaintiff or one employer. It reflects a labor market where technology is no longer just assisting the hiring process; it is shaping the candidate experience itself. If the system works well, the process feels efficient and modern. If it works badly, applicants feel invisible, employers inherit legal exposure, and everyone starts pretending they fully understand what the software did. Usually, they do not.

In that sense, the SiriusXM lawsuit is part legal challenge, part cultural marker. It shows that the era of “just trust the tool” is ending. Employers are being asked to prove fairness, not merely advertise it. Candidates are demanding transparency, not just automated updates. And courts are being asked to decide how old civil-rights rules apply in a world where employment decisions may be influenced by code, models, scoring systems, and vendor logic that few people outside the engineering team can fully explain.

Final takeaway

The class action filed in the Sixth Circuit sphere against SiriusXM is one of the clearest signs yet that AI hiring litigation is moving from theory to practice. The case is still young, the allegations remain unproven, and SiriusXM has already mounted an early defense. But the broader message is already loud enough to hear without satellite radio: employers cannot outsource accountability to software, and candidates are increasingly willing to challenge opaque hiring systems in court.

If the plaintiff’s claims gain traction, the case could become another major reference point in the law of automated hiring. If the claims fail, the lawsuit will still have served a purpose by forcing a hard conversation about transparency, validation, and the limits of algorithmic convenience. Either way, the days when companies could treat hiring tech as a frictionless back-office tool are fading fast. In 2026, the résumé pile may be digital, but the legal consequences are very real.