Addressing the AI Liability Gap in Construction Contracts

Artificial intelligence is increasingly being used to perform tasks related to construction projects, including structural calculations, cost estimates, schedules, code compliance checks, and various other functions. Yet none of the industry’s standard contract templates contain provisions prescribing responsible use of AI or allocating potential liability that could result from its use. In addition, many property and casualty insurers have issued standardized AI exclusions that became effective in policies renewing on or after January 1, 2026. Therefore, it is advisable to include language in construction contracts that targets the use of artificial intelligence.

It is critical for owners, contractors, subcontractors, and design professionals to anticipate potential risks associated with the use of AI and allocate those risks appropriately. The following examples of construction contract clauses that address AI issues merit consideration:

1. Defined Terms. Define the terms “AI Technology”, “Generative AI Tools”, “AI Input” and “AI Output”.
2. Authorization and Restriction. The use of artificial intelligence by a contractor, subcontractor, or design professional is prohibited except as specifically authorized in an attached exhibit.
3. Disclosure and Labeling of AI-Generated. Any deliverable that includes AI output should be identified with a label stating “This document includes material generated or modified by artificial intelligence (insert the name of the AI tool), and it has been reviewed by the AI user for errors and inaccuracies.”
4. Human Oversight and Standard of Care. The contract should require the AI user to maintain and comply with reasonable policies and procedures for the responsible use of AI technology, including human review and oversight of all AI input and output. The contract should also provide that the use of AI technology shall not reduce or alter the standard of care otherwise applicable to the contractor, subcontractor, or design professional under this agreement.
5. Confidentiality and Restrictions on Feeding Project Data to AI. Prohibit the input of the confidential project data to any public-facing AI tools without the data owner’s written authorization.
6. Liability Allocation and Indemnity. The AI user should be responsible for and should indemnify the contracting counterparty against claims damages, losses, and expenses to the extent caused by the AI user’s use of AI technology in performing the work or preparing any deliverable, including errors, inaccuracies, or infringement in any AI output.
7. Insurance for AI Risk. Require the AI user to confirm in writing whether it’s commercial, general liability, and/or professional liability (errors and omissions) policies exclude or limit coverage for claims arising out of the use of artificial intelligence, and require the AI user to maintain coverage that does not contain a generative AI exclusion.

It is imperative that the above contract provisions be internally consistent. For example, the party who is responsible for verifying AI output in the disclosure clause should match the standard of care and indemnity obligations in order to avoid creating internal tension. This is consistent with the fundamental contract drafting principle that the party who controls a risk should be responsible for that risk. The parties to construction contracts must be aware of the potential risks associated with the use of AI and thoughtfully mitigate and allocate those risks through reasonable contract language.