Your AI use is becoming an ethics CPE topic. Here is what that means.

What Virginia actually did

The Virginia Board of Accountancy approved an AI-ethics course that counts toward the state's 2026 ethics continuing professional education requirement. That is the news, reported through the Virginia Society of CPAs. A CPA in Virginia can now satisfy part of an ongoing license obligation by studying how to use AI ethically.

Look at what the course covers, because it tells you exactly what a board considers the ethical core of AI use: data privacy, bias, accuracy, transparency, deepfakes, and applying the AICPA Code of Professional Conduct to AI use. Those are not abstract worries. They are the specific failure points where AI gets a professional into trouble.

The framing is the point. AI is no longer treated as a personal efficiency choice that lives outside your professional duties. A board has decided that using it responsibly is part of acting like a CPA, the same way confidentiality and objectivity are.

It is worth sitting with how ordinary that decision is once you look at it. Ethics CPE has always existed to keep the profession's duties current as the world changes. New rules, new fraud patterns, new technologies all eventually find their way into the ethics requirement, because the requirement is where the profession says here is what competent and responsible looks like this year. AI fitting into that slot is not a special event. It is the requirement doing exactly what it was built to do, pointed at the most consequential new tool most CPAs have adopted in years.

What boards seem to expect you to know

You can read the course outline as a checklist of what a regulator thinks a competent CPA should be able to reason about. Each item maps to a real risk.

• Data privacy: where client data goes when you paste it into a tool, what the vendor does with it, and what your confidentiality duties require before any of that happens.
• Bias: that a model can produce skewed or unrepresentative output, and that you are responsible for catching it, not the tool.
• Accuracy: that AI can state something false with complete confidence, so anything material gets verified against the source before you rely on it.
• Transparency: being honest about where and how you used AI, with clients, with reviewers, and in your work papers.
• Deepfakes: that synthetic audio, images, and documents are now a fraud vector you have to account for in how you verify identity and evidence.
• The AICPA Code of Professional Conduct: the existing ethics rules still govern. AI does not get its own moral exemption. The Code applies to how you use it.

Notice that none of these are new principles. They are the duties you already hold, applied to a new tool. That is usually how a board approaches emerging technology. It does not invent fresh ethics. It asks whether your existing ethics hold up under new conditions.

That reframing is the most useful thing to take from the course outline. If you already take confidentiality seriously, the data privacy piece is not a new obligation, it is your existing obligation applied to a tool that can quietly send client information to a third party. If you already verify before you sign, the accuracy piece is the same duty applied to output that lies with a straight face. The work is not learning a new code of conduct. It is recognizing the specific new ways a familiar duty can be breached when AI is in the loop, and most of those ways are not obvious until someone points them out. That is precisely what a course like this is for.

For the national context, AICPA and CIMA offer a course titled "Ethics in the World of AI: A CPA's Guide to Managing the Risks." The existence of that offering, alongside Virginia's move, tells you the profession's standard setters are already organized around this material.

Do not overclaim, and do not ignore it either

Here is the balance to strike. Virginia is one state. There is no nationwide AI-ethics CPE mandate, and anyone telling you every CPA in the country now has to take an AI-ethics course is getting ahead of the facts. Requirements vary by jurisdiction, and you should confirm your own state's rules rather than assume.

At the same time, do not wave this off as a local quirk. When a state board and the AICPA and CIMA are both putting AI ethics into formal CPE channels, that is not noise. That is the early shape of a trend. The realistic expectation is that more boards follow, on their own timelines, in the next few cycles.

The practical posture is the calm one. Treat AI competence as something worth understanding for your license, not because your state has acted yet, but because the direction is clear and the material is genuinely useful. Getting ahead of a requirement is always easier than scrambling to meet it.

There is a second reason not to wait on your own board. The risks the Virginia course names are live whether or not your state has codified them. A confidentiality breach through a careless prompt is a problem the day it happens, not the day a board adds it to the CPE list. A client misled by confident, wrong AI output is exposed now. The requirement, when it comes, is just the profession formalizing risks that already exist in your practice. So the honest reason to learn this material is not compliance. Compliance is the lagging indicator. The leading reason is that the failure modes are real today, and understanding them protects your clients and your license well before any board makes it mandatory.

How this connects to the rest of your AI use

This is distinct from generic AI governance, the policies and committees a firm sets up. This is about you and your license: the ethics knowledge a board expects an individual CPA to carry. The good news is that it overlaps almost entirely with what makes you effective.

The same privacy discipline that satisfies an ethics requirement is what keeps client data out of the wrong tool. The same accuracy habit is what stops a confident, wrong AI output from reaching a client. So the ethics material is not a separate burden bolted onto your real work. It is the foundation of doing the real work well. For the hands on side, the AI tools accountants actually use covers what to reach for, and the CPA close-week AI system shows the workflow those habits protect.

This is the part most coverage of AI ethics misses. It treats the requirement as a tax on using AI, a box to check before you are allowed to get back to the productive part. That gets it backwards. A CPA who genuinely understands the privacy, accuracy, and transparency risks is also the CPA who uses AI faster, because they are not second guessing every output or quietly worrying about what they pasted where. Confidence comes from competence, not from ignoring the risks. The professionals who get the most out of these tools are usually the ones who took the time to understand exactly where the tools can hurt them.

The skill under the requirement

Whether or not your state has acted, the underlying skill is the same: using AI in a way you could defend to a board, a client, or a peer reviewer. That means knowing where AI helps, where its output cannot be trusted without checking, and how to keep your own judgment and your confidentiality duties in front of the tool at all times.

That skill is what an ethics requirement is really testing for, and it is what makes you better at the work regardless. If you want it structured, Claude AI for Accountants teaches that method for working CPAs and finance professionals, and the two minute course quiz will point you to the right starting place.