Army AI hallucination suit targets $450M White Sands award
No FAR rule governs AI in source selection, but a courtroom is about to force the issue anyway, and discovery is where the real disclosures will come.
TL;DR
TRAX International is suing in the Court of Federal Claims over a $450M Army support contract at White Sands Missile Range, alleging the source selection evaluation board relied on AI that hallucinated a weakness into TRAX's bid. GAO called the error harmless; TRAX says the Army won't disclose whether the SSEB used AI at all. The case is the first bid protest centered on a documented AI hallucination in proposal evaluation. If it survives the Army's motion to dismiss, discovery could surface how agencies actually use AI in procurements, an area with zero governing FAR rules.
The lawsuit lands at a peculiar moment. Agencies are deploying AI in acquisition workflows at speed (GSA's CALI tool does compliance checks, the IRS has automated clause review, and the Army itself has acknowledged using AI in contract document drafting) but no specific FAR rule addresses whether, how, or under what documentation standard AI can shape a technical evaluation. TRAX's complaint exploits that vacuum.
The Army's position, per the lawsuit, is that the Source Selection Authority didn't use AI to evaluate proposals. But TRAX says the Army "refused to say" whether the Source Selection Evaluation Board did. That distinction matters because, in this procurement, the SSA effectively rubber-stamped the SSEB's findings, so if AI output flowed into the SSEB report unchecked, it flowed into the award decision.
What the GAO record already shows
GAO acknowledged the Army assigned TRAX a weakness that referenced proposal material the company says doesn't exist. TRAX calls it "a classic AI hallucination, with made-up references to TRAX's proposal." GAO deemed the error harmless and denied the protest. The Court of Federal Claims isn't bound by GAO's findings, and TRAX is betting the administrative record looks different under judicial scrutiny.
The liability gap
The litigation exposes a structural problem for both sides. For contractors, proving "arbitrary and capricious" decision-making is hard when you can't see whether a human or an algorithm generated the finding you're challenging. For agencies, the absence of a rule cuts both ways: no FAR provision says using AI is improper, but no provision says it's proper either. Ira Rushing of Tully Rinckey told Federal News Network the practical standard now shaping up is that "agencies have to prove a human meaningfully reviewed and stood behind whatever AI produced." If that review isn't well documented, the agency looks arbitrary by default.
What changes Monday
For contractors bidding on federal work, the immediate takeaway is to start asking (in Q&A periods, in debriefing requests) whether the agency used AI during evaluation and, if so, at which stage. Alan Chvotkin of Protorae Law told FNN that agencies "will probably tell you a little bit, but not a lot." That may change fast if this case survives dismissal and discovery forces the Army to turn over AI-related materials.
Sen. Elizabeth Warren has already pressed DoD and seven tech firms to disclose AI contract terms, writing in July that "it's impossible to assess any safeguards" without seeing the full agreements. The TRAX case takes that transparency demand from the oversight hearing room into a courtroom where the government can be compelled to produce documents. Discovery is where the real news will break, if the motion to dismiss fails.
For now, the case is an early marker. AI is in the evaluation pipeline. No regulation governs it. A lawsuit has arrived before the rulemaking did. That sequence alone tells you where the pressure is building.
Published ·Deep Fathom