Employment litigation has quietly become the fastest-moving exposure on corporate counsel’s list, and the driver is not a new statute. It is the AI systems HR departments have folded into hiring and workforce decisions, according to a midyear survey of 135 general counsel that shows AI and workforce-change disputes climbing faster than firms expected just months ago.
The data behind the shift
Norton Rose Fulbright’s 2026 Annual Litigation Trends Survey, a midyear pulse of general counsel and in-house litigation leaders at U.S. energy, financial-services, healthcare, and technology companies, found that 47 percent now see workforce changes such as layoffs as a likely trigger for class action litigation in 2026. AI-related disputes are rising alongside them: 46 percent of respondents reported increased federal AI-litigation exposure by midyear and 42 percent reported state-level increases, figures the firm says outpaced what companies expected when it last surveyed the group in late 2025.
The exposure is concentrated, not diffuse. Employment disputes are now more likely to be litigated at the state level than the federal level, with 44 percent of respondents citing state-level increases against 39 percent at the federal level, a shift Norton Rose Fulbright attributes to compliance complexity from new employment requirements in states such as California and New York. Bias or discrimination claims tied to AI use ranked among the leading categories of anticipated litigation, alongside privacy violations and regulatory investigations. “What stands out is how quickly the litigation environment is evolving, especially around cybersecurity incidents, consumer claims and AI-related disputes,” said Steven Jansma, the firm’s U.S. head of litigation and disputes.
Where the exposure concentrates by industry
The survey’s sector breakdown shows the risk is not evenly distributed. In energy, 57 percent of respondents reported increased federal employment and labor dispute exposure, and the same share flagged workforce changes as a likely 2026 class action trigger, the highest of any sector surveyed. In healthcare, 53 percent cited increased federal AI-litigation exposure, with the same proportion pointing to AI-enabled product deployments and launches as the likely trigger. In technology, three-quarters of respondents reported increased federal litigation exposure and 72 percent reported state-level increases, with 56 percent expecting privacy or data-protection violations tied to AI use to contribute, and half citing AI bias or discrimination claims specifically.
For HR functions, the throughline across every sector is the same: AI systems that touch employment decisions, whether screening resumes, scoring performance, or selecting layoffs, are treated by respondents as a distinct and growing litigation category, separate from the broader cybersecurity and data-privacy risk that still tops the overall list. That distinction matters because it means the fix is not a general AI-governance policy. It is a specific audit trail for every employment-adjacent model.
What that risk looks like inside a company
A federal lawsuit filed this month against Meta shows what that exposure looks like once it reaches a courtroom rather than a survey response. Twenty-six current and former employees allege the company used internal AI scoring systems, drawing on performance ratings, calibration scores, and productivity metrics, to select staff for a May reduction in force, and that the scoring disproportionately caught workers who had taken or requested protected medical or family leave. The complaint details three representative plaintiffs: a scientist selected for layoff while on pre-birth pregnancy leave, a manager demoted then laid off weeks into a second medical leave, and an engineer whose rating was lowered over “broken time” tied to a work-related injury. The model’s inputs, per the complaint, included performance ratings, calibration scores, productivity and output metrics, “AI-native” ratings, and AI-token consumption, categories that penalize any employee who was, for legally protected reasons, not producing output during the measurement window. It argues the company never built in a step to neutralize those inputs for employees who could not accumulate normal output while out on leave.
That is precisely the pattern the Norton Rose Fulbright data flags as a growing category: workforce-reduction tools and AI-assisted employment decisions treated as internal efficiency systems rather than as inputs a plaintiff’s attorney can later subpoena and dissect metric by metric. The same dynamic runs through AI hiring tools that vendors marketed as a fix for human bias, where the automation itself has become the thing plaintiffs now scrutinize.
What it means for the HR leader
The practical shift is where liability now concentrates: not in the decision to use AI in hiring or workforce planning, which most large employers have already normalized, but in whether the inputs feeding those systems were audited for protected-class effects before deployment. A recruiting scorecard or a layoff-selection model that ingests raw productivity or engagement metrics, without excluding or adjusting for FMLA, ADA, or pregnancy-related leave, is exactly the kind of system the survey’s respondents now flag as a class action trigger, and exactly the kind of system a plaintiff’s attorney can reconstruct from HR’s own audit logs.
State-level exposure also means a single national policy is no longer sufficient. HR and legal teams operating across California, New York, and similar jurisdictions need AI-specific bias audits calibrated to each state’s employment requirements, not a single federal-standard sign-off applied everywhere.
What to do now
Three moves close the gap the survey and the Meta case both point to. First, inventory every AI-assisted tool touching hiring, performance, or workforce-reduction decisions, and document what inputs feed each one. Second, build and log a neutralization step for leave-affected metrics before the next reduction in force, not after litigation names it as missing. Third, treat state employment-law updates as an AI-governance trigger, not just a policy update, since the survey shows state-level exposure now exceeds federal exposure in employment disputes. Litigation risk, in this data, is no longer a downstream legal problem. It is a design requirement for the HR platforms already running the decision.
Source: Norton Rose Fulbright
