In a development that validates the exact risks our panel highlighted last December at the 2025 Construction Super Conference in Bonita Springs, Florida, Fluor Federal Services Inc. is currently arguing in Texas federal court that a subcontractor’s brief — containing fabricated quotations and misleading case characterizations generated by AI — should not be allowed to be corrected through amendment or a sur-reply.

As reported by Law360 on June 18 in A2K Inc. v. Fluor Corp. et al. (N.D. Tex. No. 3:26-cv-00206), the plaintiff staffing subcontractor admitted using generative AI tools for drafting, researching, and cite-checking its opposition to Fluor’s motion to dismiss in an antitrust case involving disaster relief contracts for the New Mexico wildfires. The brief allegedly included a non-existent quote attributed to a Texas Supreme Court decision. Fluor contends that granting a “do-over” would reward AI reliance rather than proper diligence.

AI is transforming the construction project lifecycle

This is not hypothetical. At the 2025 Construction Super Conference, our panel (Lauren Abeyta of Construction Discovery Experts, David Ehrlich of Coastal Construction, Tamara Lindsay of MasTec, Jason Loring of Jones Walker LLP, and myself) examined how generative AI is transforming every stage of the construction project lifecycle, from bidding and quantity takeoffs, real-time project monitoring with drones and sensors, to change order and claims documentation, ESI collection in disputes, and even emerging AI tools in arbitration.

A central warning we delivered then: generative AI has a “hallucination” problem (confidently producing plausible but false information, including fake citations). The panel cited the Mata v. Avianca sanctions case as an early red flag and stressed that without strong governance, human oversight, and verification protocols, the efficiency gains can quickly turn into professional liability.

Constructions lawyers must consider three things

The Fluor matter brings our panel’s Super Conference warnings into the courtroom. It underscores urgent practical questions for construction lawyers and in-house counsel:

• How do we safely integrate AI into contract review, bidding, claims preparation, and dispute work without compromising accuracy or credibility?

• What engagement letter language, billing guidelines, and quality-control processes should firms adopt?

• How do we manage privilege, ESI preservation, and potential adverse-inference risks when AI is (or is not) used on a project?

The panel’s key takeaways remain highly relevant six months later. Construction firms and their counsel who treat AI governance as optional are now seeing the consequences play out in real time.