Laurie Choi and Christian Fernandez | Snell & Wilmer
Like many industries, the construction industry is rapidly adopting artificial intelligence. AI-powered tools now help construction teams estimate costs, transcribe meetings, summarize RFIs, model schedules, and manage project documents faster than ever before. But amid this rush toward efficiency, a serious risk hides in plain sight: every time project documents are uploaded into a third-party AI tool, contractual confidentiality obligations may be implicated — or outright breached — exposing companies to potentially substantial and uncapped liability.
General contractors often face strict confidentiality and nondisclosure obligations, whether under the prime construction contract or a separate NDA. This is especially true for complex projects or those involving proprietary technology — such as data centers, medical or scientific facilities, and energy generation and storage facilities. These confidentiality clauses typically define “confidential information” broadly to encompass all non-public information provided for the project, including plans, specifications, pricing, schedules, RFPs, and even the contract terms themselves. These confidentiality obligations almost always flow down to subcontractors.
Compounding the risk, many sophisticated prime construction contracts carve out confidentiality breaches from the owner’s waiver of consequential damages. The potential consequence: when an owner suffers damages from the inadvertent release of proprietary information or trade secrets due to a contractor’s confidentiality breach, the resulting losses can snowball quickly.
Consider this scenario: A subcontractor working on a data center project for one AI model provider uses a competing provider’s platform to summarize meeting minutes, perform takeoffs, and draft supplier agreements. That action may breach confidentiality terms and could create significant trade secret exposure — particularly given the fast-evolving and competitive nature of this technology sector.
In this scenario, the general contractor faces a cascading problem. If the owner discovers the breach, the general contractor may be liable under the prime contract. The general contractor may then seek indemnification from the subcontractor, but if the subcontractor lacks the financial resources to satisfy the claim — or if its insurance policy excludes intentional data disclosures — the general contractor could be left holding the bag. Standard commercial general liability policies may not cover this type of risk, and professional liability policies may not be triggered either.
Recognizing this exposure, general contractors are now inserting express provisions into their subcontracts that prohibit the use of AI tools on project work without prior written consent. These clauses typically require subcontractors to identify which AI tools they intend to use, demonstrate that those tools do not retain or train on project data, and obtain affirmative approval before proceeding.
This creates real tension. AI has become embedded in everyday construction technology. Estimating software uses machine learning. Project management platforms incorporate AI-powered analytics. Meeting tools auto-transcribe and summarize discussions. Even email platforms now offer AI-generated responses that may process confidential content. A blanket prohibition on AI use may be technically difficult — if not impossible — to enforce.
Some contractors are getting ahead of this issue by implementing internal AI use policies. These policies typically prohibit the use of public-facing AI tools for estimating, design work, and project management, while permitting enterprise-grade platforms with contractual data-isolation guarantees. But the vast majority of subcontractors have not yet grappled with this risk. They may be violating their confidentiality obligations daily without realizing it.
While confidentiality may be the most immediate concern, AI use in construction raises several other contractual risks that industry participants should monitor. One key issue is intellectual property ownership. Architects and engineers often retain copyright in their design documents under standard form agreements, and licenses are typically limited to use on a specific project. Uploading plans or BIM models into an AI platform — particularly one that generates derivative outputs — may exceed the scope of the permitted license and could constitute copyright infringement.
Practical Steps to Protect Yourself
Whether you are an owner, general contractor, subcontractor, or design professional, the time to address this risk is now — before a breach occurs. Consider having an attorney review your current and contemplated agreements, as well as your internal AI use policies, to ensure consistency and compliance.
When one of your cases is in need of a construction expert, estimates, insurance appraisal or umpire services in defect or insurance disputes – please call Advise & Consult, Inc. at 801.641.8304, or email experts@adviseandconsult.net.