The SEC withdrew its prescriptive AI rule in 2025, then immediately placed artificial intelligence at the top of its 2026 examination priorities. For registered investment advisers operating in real estate, the message is blunt: the absence of a dedicated AI rulebook does not mean the absence of liability. The tools that accelerate deal analysis, property valuation, and client communications can also accelerate errors. And examiners will hold firms accountable under existing regulations they have already agreed to follow.
Quinn Edwards reported on July 13, 2026, that the agency's dual move creates a compliance paradox. Firms are being told to innovate without a specific AI compliance framework, yet they remain fully exposed to the broad fiduciary duties and books-and-records rules already on the books. If an AI-driven valuation model systematically overprices assets in a client portfolio, the firm cannot argue that no rule explicitly prohibited the algorithm's use.
The enforcement gap is smaller than it looks
Regulators do not need a rule titled "Artificial Intelligence" to bring an action. The Investment Advisers Act of 1940 already requires advisers to act in their clients' best interest and maintain accurate records. An AI tool that produces misleading property cash-flow projections or fails to disclose material assumptions can trigger a violation under these long-standing obligations. The SEC made clear that its examiners will ask how firms supervise AI outputs, not whether the firm used a particular vendor.
For real estate investment advisers, this means due diligence on third-party AI tools becomes a compliance function, not just an operational one. A model trained on outdated cap rate data or suburban office trends from 2019 may generate recommendations that fall short of the fiduciary standard. The firm is responsible for that gap, even if the vendor's marketing materials promised accuracy.
Where examiners will look first
The 2026 exam priorities signal three areas of focus. First, examiners will review how advisers test AI models before deployment and monitor them afterward. Second, they will scrutinize disclosures to clients about where and how AI influences investment decisions. Third, they will check whether firms maintain records of AI-generated advice in the same way they preserve traditional analyst notes. A deleted chatbot transcript is still a record.
Real estate firms often rely on AI for property screening, rent forecasting, and construction cost estimation. Each of these applications touches client assets. An error in a cost estimate that flows into a fund's net asset value calculation is not a technology problem in the SEC's view. It is a compliance problem that the firm owns.
Practical steps that reduce exposure
Firms should document their AI governance before the examiners arrive. A written policy covering vendor selection, output validation, and human override procedures signals that the adviser treats AI as a supervised tool, not an autonomous decision-maker. Testing models against real portfolio data and tracking the frequency of human overrides creates an audit trail that examiners expect to see.
Training matters. Portfolio managers and analysts who use AI tools daily need to understand the limits of those tools. A model that excels at predicting suburban single-family rents may fail on urban multifamily properties. Professionals who recognize these boundaries can intervene before a flawed output reaches a client recommendation. For teams building these skills, structured learning paths exist, including an AI Learning Path for Real Estate Brokers that addresses practical deployment and risk management.
Why this matters for real estate and construction professionals
Real estate investment advisers cannot outsource their fiduciary duty to an algorithm. The SEC's 2026 exam priorities confirm that the agency will test AI governance under existing rules, not wait for new ones. Firms that treat AI output as a starting point requiring human judgment will fare better than those that automate decisions without documented oversight. The practical takeaway is straightforward: every AI tool in your workflow needs a named person responsible for its output, a testing protocol, and a paper trail. Start building those three things now, before the exam letter arrives. For broader context on how the industry is adapting, explore AI for Real Estate & Construction resources that address compliance alongside efficiency.
Your membership also unlocks:
