By A federal judge in Michigan dismissed a lawsuit accusing Sirius XM of using artificial intelligence to discriminate against a Black job applicant, according to HRMorning. The ruling, issued without prejudice on Sept. 30, 2026, means plaintiff Arshon Harper could potentially refile his complaint.
Harper, an IT professional from Detroit, alleged that Sirius XM’s iCIMS applicant tracking system used AI features that scored candidates using stand-ins for race, including schools attended, work history and ZIP codes. He applied for roughly 150 positions between November 2023 and November 2024, received one 30-minute interview, and was rejected afterward. Harper sued under Title VII and Section 1981 and sought class action status. The court granted Sirius XM’s motion for judgment on the pleadings.
Why the court dismissed the case
The dismissal turned on pleading standards, not on whether AI hiring tools are lawful. The court assumed for purposes of the ruling that Sirius XM did use AI in hiring, then found Harper’s complaint failed on both legal theories.
For intentional discrimination under Title VII and Section 1981, Harper alleged no facts suggesting race motivated the rejections. He did not claim the job he interviewed for went to a non-Black candidate, nor did he identify similarly qualified applicants of other races who were treated better. The court’s central observation, as quoted by HRMorning: “data collection is not the same thing as data discrimination.”
For disparate impact, Harper supplied no group-level statistics, no comparison group, and no examples of other applicants with similar experiences. A disparate impact claim requires facts showing that a specific employment practice disproportionately disadvantaged Black applicants. Alleging that an AI system collects data that could correlate with race does not, by itself, establish that the system caused a discriminatory effect.
The court also noted that collecting demographic data can help employers ensure their hiring practices do not disadvantage protected groups.
How this differs from Mobley v. Workday
The Sirius XM ruling contrasts sharply with Mobley v. Workday, a parallel case that has progressed further. In Mobley, a federal court in California granted preliminary certification of a collective action under the Age Discrimination in Employment Act for applicants aged 40 and over who applied through Workday’s platform from September 24, 2020 onward and were denied employment recommendations. The Ninth Circuit found the plaintiffs sufficiently alleged a unified policy, Workday’s AI tools, as the common source of discriminatory impact.
The key distinction is the defendant. Harper sued the employer, Sirius XM. Mobley sued the AI vendor, Workday. A vendor that supplies screening tools to many employers can be challenged as the common source of a discriminatory policy across a workforce. An individual employer requires more specific facts connecting its own hiring decisions to a protected characteristic.
This does not mean employer defendants are insulated from AI bias claims. It means plaintiffs must plead facts showing how the tool’s operation caused a particular adverse decision, not merely that the tool collected data points that could serve as proxies for race.
What the dismissal does and does not mean
The dismissal is without prejudice, so Harper may amend and refile if he can allege additional facts. The ruling also does not address whether AI hiring tools can violate anti-discrimination laws. It addresses only whether this complaint stated a claim.
Employers should not read the decision as a green light for unchecked AI adoption. The Equal Employment Opportunity Commission has issued guidance on algorithmic fairness and disparate impact under Title VII. New York City’s Local Law 144 requires bias audits for automated employment decision tools. Other jurisdictions are considering similar measures. Those obligations remain in force regardless of this ruling.
The practical takeaway for HR and hiring teams is that AI tools create records. Every automated screening decision, every scored application, every rejected candidate leaves a data trail. A plaintiff who can obtain discovery into how a tool scored applicants across a protected class may be able to plead the statistical showing that Harper lacked. Employers that use AI screening should document how the tool works, what data it uses, and how human reviewers interact with its recommendations.
Steps for employers using AI hiring tools
Employers should treat this dismissal as a reminder to prepare for the next case, not as a reason to stop worrying. Specific steps include:
- Ask vendors whether their tools use proxies for protected characteristics, such as ZIP codes, school names, or employment gaps.
- Request documentation of any bias testing or validation the vendor has performed.
- Review whether human decision-makers can override automated recommendations and whether they actually do so.
- Retain records of how AI tools scored applicants, especially for positions where adverse impact claims are plausible.
The legal standard for disparate impact requires a specific employment practice, not a general allegation that AI is biased. Employers that can identify exactly which practice is being challenged will be better positioned to defend it.
Sources: HRMorning
Photo: Courtroom One Gavel - Flickr - Joe Gratz by Joe Gratz, CC0
