AI is already part of architectural practice. Firms use it to draft project specifications, create renderings, conduct research, review contracts, and handle administrative work. The central question is no longer whether firms will use AI. It is how they will review its output, document their decisions, and manage the related risks.
In the AIA Trust Week webinar, “AI in Practice: Where the Real Risk Lies,” AIA Contract Documents Chief Legal Officer Nate Budde explained where exposure appears in daily practice and what firms can do now. His message was practical: AI may accelerate the work, but it does not replace professional judgment or accountability.
Watch the Webinar Replay
Where AI Risk Enters Architectural Practice
AI-related exposure does not begin with futuristic scenarios. Nate discussed familiar uses, including drafting specifications, generating concept renderings, and using general-purpose AI to summarize contract language. Each tool can produce a plausible result quickly. Risk enters when that result influences a professional deliverable or contract decision without appropriate review.
A firm may use AI for early exploration and internal drafting. The stakes rise when the output becomes part of a drawing, specification, client recommendation, or signed agreement. A polished answer is not necessarily an accurate or verified answer.
“The real question is whether you or the people in charge of your firm know how it’s being used.”
Nate Budde, Chief Legal Officer, ACD
The Standard of Care Still Applies
Nate described AI as a new production method, not an exception to an architect’s professional obligations. If a dispute arises, reviewers may examine the project record long after someone entered the prompt. That record should show what the firm checked, who reviewed the work, and why the final result was reasonable.
“The tool may change how much time you have, but it doesn’t change the expected result.”
The responsible professional must still be able to explain and defend the final work. Speed and convenience do not establish that a deliverable meets the standard of care.
Review AI-Related Contract Language Before Signing
Risk can enter through contract terms as well as design work. AI-specific clauses may address disclosure, permitted use, confidentiality, ownership, indemnification, limitations of liability, and insurance. Terms that promise error-free outcomes or raise the firm’s ordinary professional obligations can create exposure beyond the underlying agreement.
Pay particular attention to provisions that turn a duty to exercise professional care into a guarantee of results. Confirm that the contract’s expectations align with the firm’s review process and available insurance coverage. Ask legal and insurance advisers to evaluate project-specific language when appropriate.
Quick Decision Guide: AI Tasks and Review Controls
Use the level of review that matches how the AI-assisted work could affect the project, the client, or the firm’s contractual obligations.
AI-assisted task | Primary question | Practical control |
Drafting specifications | Are references, requirements, and coordination correct? | Technical review by the responsible professional |
Concept renderings | Could the image create an inaccurate client expectation? | Check the image against design intent and identify it as conceptual |
Contract summaries | Did the summary omit critical exceptions or obligations? | Compare it with the full agreement and obtain legal review as needed |
Client-facing deliverables | Can the firm explain and defend the final output? | Assign a named reviewer and retain the approval record |
Build a Written AI Use Policy Your Team Can Follow
Staff may already be using AI in ways the firm’s existing policies do not address. Nate explained why that gap is emerging.
“It’s because adoption at the individual level is outpacing the firm’s governance and controls at the group level.”
An initial AI policy does not need to be long. A clear one- or two-page policy can establish useful guardrails if staff understand and follow it. Identify approved tools and the person responsible for maintaining that list. Define permitted tasks, information-handling rules, review responsibilities, and the evidence that belongs in the project record.
The policy should also explain how staff requests approval for new tools, when they should escalate questions, and how the firm communicates AI use to clients. The goal is a repeatable process that fits existing project workflows.
Discuss AI Expectations With Clients Early
Address AI use during proposal and contract discussions, before an unexpected disclosure undermines trust. Align expectations about where the firm may use AI, how qualified professionals will review the work, and how the firm will protect client and project information. The conversation should reflect practices the firm can follow consistently.
What Architecture Firms Can Do Next
Start with the AI use already happening inside the firm, then add controls that match the risk of each task.
- Identify where staff use AI in project and administrative workflows.
- Separate low-stakes drafting assistance from work that affects project deliverables or contract decisions.
- Name qualified reviewers and define the verification the project record must show.
- Review AI-related contract terms for changes to the standard of care, liability, and insurance expectations.
- Adopt a concise written policy, train the team, and update the policy as tools and practices change.
Nate’s message was not to stop using AI. It was to make AI use visible, reviewable, and defensible. Firms that establish practical controls now will be better prepared to use these tools while maintaining professional responsibility and client trust.
Watch AI in Practice: Where the Real Risk Lies
Hear Nate’s full discussion of practical AI exposure, professional standards, contract provisions, and the controls architecture firms can put in place now.
