The Interactive Advertising Bureau released version two of its AI Transparency and Disclosure Framework, aiming to establish an industry baseline for when and how AI use in ads should be disclosed.
Since the framework was first published in January, laws and regulations surrounding AI disclosure have taken effect in the European Union, Asia, New York, and California — three continents now requiring AI labeling in some capacity.
Version two does not change the previous recommendations. It expands them to incorporate the fast-changing regulatory landscape, adding how marketers can use the labels required by law.
The most operationally useful part of the update is the recommendation to balance transparency against business objectives. The framework warns that overusing AI disclosures leads to "label fatigue," making labels less effective long term.
Caroline Giegerich, vice president of AI at the IAB, summarized the position:
"Trust is everything between a brand and its customers, and being honest about AI is part of earning it. That said, not every use of AI needs a label — labeling everything teaches consumers to ignore labels and could negatively impact advertisers. This is why we take a meticulously nuanced position in this framework."
Version two provides evidence-based support for the guidelines.
Consumer data is mixed. Data in the framework found 73% of millennial and Gen Z consumers said an AI-generated ad would have no effect on, or even increase, their likelihood of purchase. Other studies have found AI-generated ads can negatively impact click-through rates.
Disclosure is a compliance question and a brand question. Meet the legal minimum, then decide separately whether to disclose more as a trust strategy. The IAB's argument is that over-disclosure erodes the informational value of the label itself.
Sort by consumer perception, not by technique. Color correction may use AI without any risk of consumer misapprehension. Synthetic people and synthetic voices are different. Writing that line into internal guidelines lowers the cost of case-by-case judgment.
Regulation has already fragmented by region. The EU, Asia, New York, and California each impose different requirements. For global campaigns, building creative to the strictest market's standard and swapping only the regional label is cheaper to manage.
Assume AI is already deep in the production pipeline. The observation that AI now runs the entire short-form pipeline, making trust rather than production the scarce input is the backdrop to this framework. Industry self-regulation moving ahead of law also appeared in ByteDance's IP protection deal with the Motion Picture Association.
The IAB's message reduces to two sentences. Being honest about AI is part of earning trust. But labeling everything makes labels meaningless. The practical work is documenting your own line between them.
It keeps the prior recommendations and expands them to reflect regulations now in effect in the EU, Asia, New York, and California, including how marketers can use legally required labels.
Consumer-facing cases: synthetic images or video, digital twins of living and deceased persons, synthetic voices in specific situations, and conversational agents used in advertising.
Standard editing such as color correction and other post-production work, plus clearly stylized or fantastical imagery.
When overusing AI disclosures teaches consumers to ignore labels, reducing their long-term effectiveness — which the IAB warns could negatively impact advertisers.
To apply what you just read to your own site, start with a free audit of where things are now.
A strategist replies within 24 hours on business days.
In CreatorIQ's survey of 5,095 creators, suitability topped the creator selection criteria and follower count came dead last. Yet follower count still predicts pay better than anything else.
Brand equity comes from accumulation, not novelty. How a Korean travel platform kept the same jingle for six years without stagnating.
Chipotle turned a 16-year-old YouTube star's order into a limited Kid's Meal, extending a custom-order tactic that worked on Gen Z to a younger cohort.