The Interactive Advertising Bureau released an update to its AI Transparency and Disclosure Framework on Tuesday, responding to new disclosure laws in New York, California, South Korea, and the European Union. The guidance is the first revision since January and aims to give advertisers a practical path through a regulatory landscape that is shifting faster than most compliance teams can track.
The core recommendation: targeted disclosure rather than universal labeling. The IAB suggests advertisers label AI use only when it "materially affects authenticity, identity, or representation in ways that could mislead consumers." In plain terms, the decision to slap a sparkle icon on an ad should depend on how the AI was used, not on whether AI was used at all.
When to disclose
The IAB recommends disclosure for AI-prompted images and videos, certain synthetic voices, avatars, digital twins of deceased or living individuals, and AI chatbots or assistants. The organization suggests advertisers can use a "standardized sparkle icon or text label" to flag these uses.
Caroline Giegerich, VP of AI at the IAB, acknowledged the difficulty of keeping pace with new laws. "There's been a lot of legislation that has matured, come into effect," she told Marketing Brew. "Even doing the legislation section for this Part 2, I'm constantly thinking: 'Am I missing something new?' And I know that it'll be out of date by literally one day, but this is just kind of the environment we're living in…I still think [the framework update] has an intention as the voice of the industry."
When you can skip the label
The IAB also carved out exceptions where disclosure likely isn't necessary: routine postproduction, internal workflows, clearly fantastical imagery, authorized synthetic voices of real people, generic synthetic voices, background music, standard audio enhancement, digital twins in standard brand endorsements, obvious cartoon or stylized avatars, and text or copy.
"Our overall perspective is if it's not labeled, and it causes consumer deception," then that could be a problem, Giegerich said. She offered an example: a consumer buys an under-eye cream after seeing a TikTok endorsement, then realizes the product's results were depicted using AI. That gap between expectation and reality is where regulatory risk lives.
New York State's AI disclosure law, which requires advertisers to be transparent about "synthetic performers," went into effect earlier this year. Some marketers welcomed the shift. "Consumers in this market demand transparency more than they ever have before," Johnny Rohrbach, co-founder and global head of partnerships at Silverside.AI, said at the time. "AI is very much a medium through which to tell stories, but I don't think it makes sense from a business perspective to try and hide the fact that you were using it."
Why this matters for marketers
The IAB's update gives marketing teams a workable starting point for compliance, but it's not a permanent answer. The organization itself expects the framework to need further revisions as more legislation takes effect. For now, the practical takeaway is to build a review process that asks one question per campaign: could this specific AI use mislead a reasonable consumer? If the answer is yes, label it. If the answer is no, document why - because regulators are increasingly asking for that reasoning.
For marketing professionals looking to stay current on AI practices and compliance considerations, resources like AI for Marketing and AI for Marketing Managers offer practical guidance on integrating AI tools responsibly.
Your membership also unlocks:
