The EU AI Act's August Deadline: Why US Agents Should Care

On August 2, the EU AI Act's transparency rules, known as Article 50, took effect. Most US agents don't have European clients, so it is easy to file this under "somebody else's problem."

But I'd keep it on your radar. EU rules have a way of crossing the Atlantic, and what this one requires looks a lot like where US disclosure rules for listings and advertising are already headed.

What the EU rules require

In short, AI-generated or AI-altered images, audio, and video that look like real people, places, or events have to be labeled as artificial. The label requirement applies even when nobody meant to deceive anyone. Chatbots also have to tell people they are talking to an AI.

There was some confusion in July about whether the EU's AI Omnibus, a package that took effect in late July to simplify parts of the Act, delayed all of this. It did not. Most of the transparency duties started on schedule. The one carve-out is narrow: generative AI systems already on the market before August 2 have until December 2 to add machine-readable marking to their output. Fines for non-compliance can reach 15 million euros or 3% of global annual turnover, whichever is greater.

The Omnibus was a response to complaints that the original rollout was too fast and too vague, especially for smaller companies. That shows that regulators are listening to the cost of compliance, but they are not backing away from disclosure itself.

The US is already moving the same direction

The state level is where this gets real for us in the United States:

• California has a law in effect since January 1, 2026 requiring brokers and agents to disclose when listing images have been digitally altered and to provide access to the originals. Basic edits like lighting, cropping, and color correction are carved out as long as they don't change how the property actually looks.
• Hawaii signed a deepfake law in July that requires disclosure of synthetic performers in advertising.
• Other states have bills pending that target AI-altered listing photos and deceptive AI in housing ads.

There is still no federal standard, and states are not writing identical rules. That is the part to plan for. If you work in more than one state, or your brokerage does, the safest approach is to follow the strictest rule you are likely to face.

What to do now

• Write down a policy on AI-altered listing media. It doesn't need to be complicated. Virtual staging, AI-brightened rooms, and generated twilight shots are all covered.
• Disclose in the listing itself. A short line like "virtually staged" or "image digitally altered" costs nothing and protects you from a misrepresentation claim.
• Keep the originals. If a buyer or regulator asks to see the unaltered photo, you should be able to produce it.
• Ask your vendors. If your staging, photo-editing, or marketing tools generate images, ask whether they label their output and what they plan to do as rules change.
• Check your MLS rules. Several already have AI disclosure requirements, and they tend to move faster than legislatures.

What to take from this

You obviously don't need to track every bill in every state. But the direction is clear and it's the same one the EU just made official: if AI touched it and a reasonable buyer would care, say so. Build that habit now, while it's still voluntary in most places, and a new state law will be a minor adjustment instead of a scramble.

- Jason