IAB has updated the AI Transparency and Disclosure Framework it launched in January, getting more specific about when AI in advertising should be disclosed.
The key distinction in V2: AI use itself doesn’t trigger disclosure.
IAB says disclosure should focus on AI that materially affects “authenticity, identity, or representation” in ways that could mislead consumers.
That means prompt-generated realistic images and video, certain synthetic voices, fabricated digital twins and AI assistants that could be mistaken for humans generally warrant disclosure.
But AI-generated text and copy, routine post-production, generic synthetic voices, background music, obviously stylized avatars and internal AI workflows generally do not.
Behind the consumer-facing disclosure is a second issue: provenance . IAB recommends C2PA Content Credentials and machine-readable metadata that can indicate whether disclosure is required and whether a label was applied.
Read:
- “AI Transparency and Disclosure Framework Version 2” (August 18) – IAB
- IAB Updates Industry Framework for Consistent AI Transparency & Disclosure in Advertising (August 18) – IAB
From tipsheet: In May, Mobile Dev Memo analyst Eric Seufert raised a question : would consumers “even care about the provenance of ads in the first place?”
Separately, when Google and Meta expanded AI disclosures in July, we noted the provenance infrastructure underneath them: metadata about how creative was produced and carried through the advertising supply chain.
IAB’s V2 starts to connect those ideas. Don’t disclose the use of AI. Disclose when AI materially changes what a consumer might believe is real.
The consumer doesn’t need to see the provenance of everything. Still, the advertising system may need to know it.
AGENTS
Google makes agentic case to publishers
In a Google-sponsored post on Digiday, Google proposes that Google Ad Manager, its publisher ad server, become an execution environment for agentic advertising.
Read: How publishers can strengthen connections with advertisers in the AI era (August 18) – Google on Digiday
This positioning echoes Google’s June announcement which said it is developing infrastructure allowing first- and third-party agents to interact at scale, specialized publisher and agency agents, plus REST APIs and an MCP server supporting trafficking workflows. It also explicitly described agents eventually handling inventory discovery, pricing and campaign execution. Read that one (June 18).
Within Google’s pitch to publishers is Buyer Direct which we’ve covered recently :
- “ Monetize through diverse transaction paths: From there, publishers can make their verified audience segments discoverable and actionable for buyers across different transaction types, from guaranteed and private marketplace deals to buyer direct , opening up inventory to incremental budgets.”
All of this comes as the industry awaits final remedies from Judge Leonie Brinkema at the Google ad tech monopoly trial where the company was found liable in April 2025 for monopolizing publisher ad serving.
From tipsheet: Google is telling publishers AI will give them more control while putting Google Ad Manager at the center of that future.
The timing matters: Google is making the case for GAM as open, agentic infrastructure while a federal judge decides what remedies are necessary to restore competition around that infrastructure.
The irony: the remedy question is whether Google has too much control over sell-side infrastructure, while Google’s AI strategy is to make that infrastructure the place where everyone else’s agents come to do business.
LLMs & CHATBOTS
Developments
- Introducing ChatGPT for Teens (August 18) – OpenAI
- Alibaba’s Small, On-Device Model Gains Traction (August 18) – The Information (subscription)
- Operation Blue Skies: Reducing aviation climate impact with AI (August 18) – Google
MARKETERS
Do you need that AI?
Ad Age’s Lindsey Rittenhouse writes:
- “Marketers are feeling pressure to adopt AI on many fronts, from internal factions eager to integrate cost-saving AI into workflows, to overtures from agencies eager to sell them their proprietary tools and capabilities—whether they need them or not. And that rush to adoption threatens to obscure the end goal.”
Read: How marketers can evaluate AI tools—experts weigh in on where to invest (August 18) – Ad Age (subscription)
MOBILE
Apple and tracking the consumer
Mobile Dev Memo analyst Eric Seufert commented on a recent ruling by Germany’s competition authority that requires Apple to rewrite its App Tracking Transparency (ATT) rules so that it does not favor Apple.
Seufert writes on X :
“I believe the agreed-upon changes to the ATT prompt, as well as the ability for developers to bundle other data-use consent requests with the ATT prompt, will nudge opt-in rates upward by a non-trivial amount. Apple has stated that the changes will apply in almost all EU countries, and I expect the remaining European investigations to resolve similarly. But I’m skeptical that the US will adopt similar measures, and for that reason, the broader impact on the digital advertising market of this restructuring of ATT will be muted…”
Read: “Apple forced to restructure ATT” (August 18) – Eric Seufert, analyst on Mobile Dev Memo
From tipsheet: Companies like Meta and AppLovin have spent years building ad systems that could perform with less Apple data. Now they may get some of that data back.
Related: “Performance Advertising: How we drive value and handle data” (March 31) – CEO Adam Foroughi on AppLovin blog
TECH
Humans move toward decisioning
Robbie Stein, VP Product, Google Search on X yesterday:
-
“The most important skill for product managers is now decision making. Building is no longer the barrier. You should be doing everything possible to learn the art and science of making great decisions.”
From tipsheet: AI is pushing human jobs up the stack, from managing and building to deciding what is worth doing in the first place.
MORE
- LiveRamp Shareholders Approve $2.2B Publicis Deal, But Reject Execs’ $82.6M Payday (August 18) – Adweek
- ANA cautions marketers against overreliance on retail media data (August 18) – Marketing Dive
- Taking A Look At Tuple, A New Entrant To The Ossified DSP Market (August 18) – AdExchanger
- How DTC startups are using AI to scale more efficiently (August 17) – Modern Retail
- Commercial conversations in ChatGPT more than doubled in a year (August 13) – Profound