AI Regulation Tracker / Guidance and standards

USPAP Advisory Opinion 41 Guides How Appraisers Use AI and Automated Tools

Q: If I use an AVM or AI output, who is responsible for the result?

The appraiser. AO-41 states that technology may assist an appraiser but cannot replace the appraiser's judgment or responsibility. Relying on a tool does not move USPAP accountability to the vendor or the model. The output still must be supported by the appraiser's own evidence, reasoning, and analysis.

On April 23, 2026, the Appraisal Standards Board of The Appraisal Foundation adopted Advisory Opinion 41, Use of Technology in an Appraisal or Appraisal Review Assignment. It explains how existing USPAP duties apply when an appraiser uses AVMs, statistical software, or generative AI. It is guidance, not a new rule, and the appraiser stays responsible for the result.

By Anthony Guerriero · The Leveraged Years Regulatory Research · Published July 19, 2026 · Last updated July 19, 2026

The short version

On April 23, 2026, the Appraisal Standards Board adopted Advisory Opinion 41 (AO-41), "Use of Technology in an Appraisal or Appraisal Review Assignment." It tells appraisers how their USPAP responsibilities apply when they use technological tools, including automated valuation models, regression and statistical software, and generative AI, across real, personal, and intangible property. The core message is that the appraiser stays in control and stays responsible. Leaning on a model or an AI output does not move USPAP accountability off the appraiser. Four rules do the heavy lifting: ETHICS, COMPETENCY, SCOPE OF WORK, and RECORD KEEPING. Important ceiling: an Advisory Opinion is non-binding guidance. It creates no new requirement. USPAP itself is the binding standard, enforced through federal law and state appraiser boards, and AO-41 only clarifies how that binding standard already applies to these tools.

Regulatory Briefing

Instrument: Advisory Opinion 41 (AO-41), "Use of Technology in an Appraisal or Appraisal Review Assignment"
Authority: Appraisal Standards Board (ASB) of The Appraisal Foundation, the body that authors and interprets USPAP
Jurisdiction: United States
Bindingness: Non-binding guidance. An Advisory Opinion does not establish new standards and does not interpret existing standards as if it were text of USPAP. USPAP itself is the binding standard.
What Changed: Nothing in an appraiser's underlying duties. AO-41 explains how the existing ETHICS, COMPETENCY, SCOPE OF WORK, and RECORD KEEPING rules already apply when a tool or AI is used.
Who This Affects: Real property, personal property, and intangible property appraisers, plus lenders, AMCs, and reviewers who rely on their work
Current Status: Adopted April 23, 2026; available now and to be incorporated into a future edition of the USPAP Guidance and Reference Manual
Primary Source: The Appraisal Foundation, appraisalfoundation.org
Last Verified: July 19, 2026

The Appraisal Standards Board is the group inside The Appraisal Foundation that writes and interprets USPAP, the Uniform Standards of Professional Appraisal Practice. On April 23, 2026, it adopted Advisory Opinion 41, which it titles "Use of Technology in an Appraisal or Appraisal Review Assignment." The opinion answers a question appraisers have been asking out loud for two years: what happens to my professional responsibility when a machine does part of the work.

The short answer is that nothing moves off the appraiser. AO-41 covers the full range of tools an appraiser might reach for, including automated valuation models, regression and statistical software, and generative AI, and it covers all three appraisal disciplines, real property, personal property, and intangible property. Across all of it, the appraiser is the one on the hook for the result.

What AO-41 actually says

The central theme is control. The appraiser determines the scope of work, evaluates whether a given tool is appropriate for the assignment, assesses the reliability of both the inputs and the outputs, and applies professional judgment to reach the assignment results. A tool can speed that up. It cannot own it. As AO-41 puts it, "Technology may assist an appraiser, but it cannot replace the appraiser's judgment or responsibility."

That single line is the whole point. If you feed a property into an AVM, or ask a generative model to draft a market narrative, or run a regression to support an adjustment, the output is an input to your analysis, not a substitute for it. It still has to be supported by relevant evidence, sound reasoning, and your own independent work. Reliance on an AI answer does not transfer your USPAP accountability to the vendor, the model, or the data provider.

The four rules that carry the weight

AO-41 does not invent duties. It maps existing USPAP rules onto the way these tools get used. Four rules are the ones that matter most.

The COMPETENCY RULE. This is the sharpest one for AI. You cannot competently use a tool you do not understand. AO-41 frames the standard directly: "The appraiser must have sufficient knowledge and experience to use the tool competently, recognize when its output may be unreliable, and determine whether reliance on the output is appropriate." If you do not know how a model was built, what data it was trained on, or where it tends to break, you are not in a position to judge whether its output is credible. That is a competency gap, and USPAP already addresses competency gaps.

The SCOPE OF WORK RULE. Deciding whether to use a tool at all is a scope-of-work decision. The appraiser sets it, documents it, and owns whether it produces credible assignment results.

The ETHICS RULE. Confidentiality and impartiality do not pause when a tool enters the workflow. Feeding confidential assignment data into a third-party AI system, for example, is a real ethics question, not a convenience.

The RECORD KEEPING RULE. The workfile still has to prove the work. AO-41 restates the standard plainly: "The workfile must contain sufficient documentation to support the appraiser's opinions and conclusions and show compliance with USPAP." If a tool contributed to your conclusions, your file needs to show enough about how you used it and why you trusted its output.

What this is, and what it is not

I want to be exact here, because the enforcement mechanics behind USPAP make the distinction matter.

AO-41 is an Advisory Opinion. Consistent with every Advisory Opinion, it does not establish new standards and it does not carry the binding force of USPAP text. It is guidance on how to apply rules that already exist. Nobody woke up on April 24, 2026 with a new legal duty because AO-41 was adopted.

What is binding is USPAP itself. USPAP is the recognized standard for appraisal practice in the United States, embedded in federal law through FIRREA for federally related transactions and enforced by state appraiser regulatory boards through licensing and discipline. The ETHICS, COMPETENCY, SCOPE OF WORK, and RECORD KEEPING rules that AO-41 discusses are enforceable. AO-41 is the map, not the territory. But the map tells you exactly where the enforceable lines already run when AI is in the picture, which is why it matters even though it adds nothing new.

What this means for appraisers and the people who rely on them

If you appraise real estate, personal property, or intangibles, AO-41 is your written answer to the "can I just use the AI number" question. You can use the tool. You cannot outsource the judgment or the responsibility. Before you rely on an output, you need enough understanding of the tool to know when it is wrong, a documented reason it was appropriate for this assignment, and a workfile that shows your reasoning rather than the model's.

For lenders, appraisal management companies, and reviewers, the takeaway is that an AI-assisted appraisal is still fully the appraiser's product under USPAP. The presence of a model in the workflow does not dilute the appraiser's accountability, and it does not give you a reason to relax review. If anything, the competency and documentation angles give reviewers a cleaner checklist.

And for finance and valuation professionals watching this from the accounting side, AO-41 is a clean example of where standard-setters keep landing on AI. The tool assists, the licensed professional stays responsible, competency to use the tool is a real obligation, and the file has to prove the work. That posture travels well beyond appraisal.

What to do now

Read AO-41 as a checklist, not an alarm. Confirm you understand the tools you already use well enough to catch a bad output, since that is the competency standard in practice. Write down, as a scope-of-work decision, why a tool is appropriate before you rely on it. Keep confidential assignment data out of systems that do not protect it. And make your workfile carry the proof: what the tool contributed, why you trusted it, and how your own analysis supports the result. None of that is new. AO-41 just tells you, in one place, that USPAP already expects it.

Questions professionals are asking

Does AO-41 create a new rule about AI in appraisal?

No. AO-41 is an Advisory Opinion adopted April 23, 2026. It does not establish new standards. It explains how existing USPAP rules, mainly the ETHICS, COMPETENCY, SCOPE OF WORK, and RECORD KEEPING rules, already apply when an appraiser uses technology, including AVMs, statistical software, and generative AI. USPAP itself is the binding standard.

If I use an AVM or AI output, who is responsible for the result?

The appraiser. AO-41 states that "Technology may assist an appraiser, but it cannot replace the appraiser's judgment or responsibility." Relying on a tool does not move USPAP accountability to the vendor or the model. The output still must be supported by the appraiser's own evidence, reasoning, and analysis.

What does the COMPETENCY RULE require when using AI?

AO-41 says the appraiser must have sufficient knowledge and experience to use the tool competently, recognize when its output may be unreliable, and determine whether reliance on the output is appropriate. In practice, if you do not understand how a tool works well enough to catch a bad answer, you are not yet competent to rely on it for that assignment.

What property types does AO-41 cover?

All three appraisal disciplines: real property, personal property, and intangible property. The guidance applies across appraisal and appraisal review assignments regardless of the asset type being valued.

Is AO-41 enforceable against appraisers?

Not on its own. It is non-binding guidance. What is enforceable is USPAP, which is embedded in federal law through FIRREA for federally related transactions and enforced by state appraiser regulatory boards. AO-41 clarifies how those binding rules apply to AI and other tools.

RELATED BRIEFINGS

Browse the full AI Regulation News tracker

Informational analysis for working professionals, not legal, appraisal, or compliance advice. Confirm how USPAP and any Advisory Opinion apply to your assignment with qualified professionals and your state appraiser board.