The IAB is entering the AI accountability discussion with a new framework aimed squarely at one of online marketers’ greatest open inquiries: when, exactly, should AI utilize be revealed in advertising?

On Thursday, the trade group rolled out its initial AI Transparency and Disclosure Framework, positioning it as a useful guide for brands, companies, publishers and systems browsing generative AI at scale.

As opposed to imposing covering disclosure guidelines, the framework adopts a risk-based technique that concentrates on consumer impact– divulging AI utilize only when it materially impacts credibility, identification or depiction in ways that can deceive people.

Dig deeper: In 2026, human connection becomes advertising and marketing’s actual advantage

“We are absolutely at a vital inflection point with generative AI,” David Cohen, Chief Executive Officer of IAB, stated in a statement. “While AI is transforming how we function from ideation to execution and dimension, we need to obtain transparency and disclosure right, or we risk shedding the count on that underpins the entire worth exchange.”

At the heart of the structure is a basic question: Does AI participation meaningfully change what a customer thinks they’re seeing, listening to or interacting with? If the response is yes, disclosure is expected. That consists of situations such as AI-generated or heavily synthesized pictures and video clips portraying real-world occasions, artificial voices of genuine people making statements they never made, electronic doubles positioned in circumstances that never ever took place, and conversational representatives or avatars designed to imitate human communication in ads.

Importantly, the IAB is not asking for disclosures every time AI is associated with a project. Routine uses– such as AI-assisted editing, optimization or background workflows– do not automatically cause labeling. The idea is to prevent disclosure overload while still safeguarding customers from being misinformed.

Dig deeper: 5 advertising and marketing lessons hiding inside Disney’s retailing engine

To make this practical across networks and systems, the structure presents a two-layer version. One layer is consumer-facing, using standardized message labels or visual signs like badges, icons, watermarks or interactive information components placed near the ad imaginative. The other layer is machine-readable, relying upon metadata standards such as C 2 to sustain technological conformity and downstream openness.

For online marketers, the framework is much less concerning inspecting a compliance box and even more about future-proofing AI fostering. As regulatory authorities, systems and customers look at AI utilize more closely, having a common industry typical provides teams an extra uncomplicated way to balance speed, imagination and obligation– without guessing where the line is.