The Rise (and Fall?) of the AI Whistleblower

Dr. Herbert Remidez, AI Professor and Serial Whistleblower

Dr. Herbert Remidez, PhD, is not your traditional whistleblower. According to his online profile, he is an Associate Professor of Business Analytics at the University of Dallas’ Gupta College of Business in Texas. His teaching areas include cloud computing, AI, and “advanced spreadsheet modeling.” In his official University bio, Dr. Remidez states that “[o]ver my 20+ years as a researcher, my research interests have included promoting trust in online discussions to improve decision- making, understanding the important dimensions of online learning platforms, improving project cost and duration estimates, and enhancing fraud detection in government-funded programs (my current research emphasis).”

Dr. Remidez appears to be profiting quite handsomely from his current research emphasis. In June 2025, for example, his company H Remidez, LLC, received $850,500, as part of a False Claims Act (FCA) settlement with Delta Air Lines. The case was triggered by his company’s filing of a whistleblower complaint under the FCA’s qui tam provisions. Delta ultimately agreed to pay $8.1 million to resolve allegations that it violated the FCA by agreeing to executive pay limitations when accepting certain COVID-era relief funds and then paying its executives in excess of those limitations.

But Dr. Remidez has no apparent ties to Delta. No insider knowledge. No smoking gun email. Instead, it appears that Dr. Remidez used his AI research skills to comb publicly available data (presumably including databases of COVID relief fund recipients and cross-referencing with publicly available SEC filings) and used those findings to bring his lawsuit. And the Delta lawsuit is just the tip of the iceberg. Dr. Remidez’ LLC is listed as the plaintiff in various other pending qui tam lawsuits throughout the country, including one against various national restaurant chains and another against a New York-based healthcare system.

Proliferation of Data Mining Qui Tams

When the FCA was signed into law by President Abraham Lincoln in March 1863, its main goal was to punish “unscrupulous contractors” who sold faulty goods such as “rancid rations” and “decrepit horses and mules in ill health” to the Union Army during the Civil War. It included a qui tam provision (a concept inherited from English common law) which allowed private citizens to bring to the government’s attention information related to such fraud, in exchange for a bounty.

Today, the FCA, and the whistleblowing that fuels it, is a multi-billion-dollar-per- year business. In fiscal year 2025, for example, the U.S. Department of Justice (DOJ) collected more than $6.8 billion in FCA settlements and judgments. The majority of that (over three-quarters) was the result of qui tam whistleblower lawsuits. In fiscal year 2025 alone, FCA relators (and their legal counsel) received a total of over $330 million, which was actually down significantly from the year before when it exceeded $479 million.

Recently, the DOJ confirmed that out of the nearly 1,300 new qui tam lawsuits filed in fiscal year 2025, over half were filed by data mining entities like H Remidez, LLC. It appears that just like every other aspect of human existence, AI bots have found their way into the world of the FCA. The use of data mining and AI in fraud detection is not new. Government agencies (including the Centers for Medicare and Medicaid Services and its contractors) as well as private health insurance companies have been utilizing data analytics to search for indications of fraud and abuse for well over a decade. Over the past several years, FCA enforcement has also increasingly been driven by data, both by government agencies and by qui tam relators. The DOJ has boasted that advances in analytics allow investigators to identify billing anomalies, outliers, and patterns suggestive of fraud. Qui tam relators and their counsel have followed suit. Rather than relying on insider knowledge—the traditional hallmark, and original purpose, of qui tam actions—complaints like the one filed by Dr. Remidez’s LLC are based largely on publicly available data, statistical modeling, or extrapolation from industry-wide trends. These cases often allege fraud without any direct evidence of misconduct by the defendant.

Presumably, qui tam actions like the one against Delta that are based solely on analysis of publicly available data should be barred by the FCA’s public disclosure bar, which provides for the dismissal of a qui tam action if “substantially the same allegations or transactions as alleged in the action or claim were publicly disclosed.” But the public disclosure bar is nuanced and gives the government the unfettered opportunity to “veto” any such dismissal. Perhaps even more relevant, before companies like Delta even have the chance to seek such relief from a court, they often have to undergo years-long investigations, which can cost millions of dollars in legal fees and expenses. Put simply, its often easier and cheaper to settle than to fight.

The DOJ Cracks Down

Against that backdrop, the DOJ appears increasingly concerned that many of these data-driven qui tams may lack sufficient factual grounding or may impose undue investigative burdens on the government, noting that cases initiated by its own internal data analysis (based on “more detailed, non-public data”) generally results in more successful outcomes than cases initiated by data-driven qui tam relators. To deal with this increasingly problematic trend, on April 30, 2026, the DOJ’s Civil Division unveiled its new FOCUS Initiative, aimed specifically at data-mined qui tam complaints.

Although the DOJ says that this initiative is intended to “materially strengthen” the government’s relationship with whistleblowers, it seems more likely to have the opposite effect. To be clear, the DOJ does not appear ready to eliminate the AI whistleblower completely. Instead, it hopes that the FOCUS Initiative will help it “prioritize working with data miners who demonstrate an insightful application of sophisticated technological capabilities to regulatory frameworks to help identify potential fraud that would otherwise go undetected.” To accomplish this goal, the DOJ will look for data miners who “provide valuable leads through high-quality, reliable, and predictive data analyses and signals and a thorough understanding of the relevant legal obligations.” Data miners should also “assess potential alternative explanations for the observed conduct and be able to articulate how the data, in combination with other available evidence, suggests both scienter and falsity.” The DOJ also encourages data miners to “partner with others who can aid their understanding of program eligibility requirements and regulatory frameworks.”

The End of the Gold Rush?

The FOCUS Initiative is not the end of the AI whistleblower, but the future is likely to be far more constrained than early success suggests. Data analytics can help uncover fraud that would otherwise remain hidden, particularly in complex industries where misconduct may be buried in thousands of claims, contracts, or public filings. But the FCA was not designed to reward speculative lawsuits built on public data alone, especially where the relator lacks insider knowledge, direct evidence, or a meaningful understanding of the regulatory framework at issue. The FOCUS Initiative signals that the government still sees value in sophisticated data- driven fraud detection, but only when it is paired with legal precision, reliable methodology, and evidence that plausibly supports falsity and scienter. In other words, AI may remain a powerful tool in FCA enforcement, but the era of filing mass qui tam complaints based on data anomalies alone may be nearing its end.

GWB represents businesses and individuals in connection with government investigations and litigation, including False Claims Act litigation. If you need assistance with such a matter, please contact us.