People no longer believe what they see, and the advertising industry has not agreed on the honesty in responding to this.
Do AI generated backgrounds need labels? What about a synthesized soundtrack? Or does the threshold only apply if it is a human face, a product claim, or a body that never existed? This feels like a trivial difference until you factor in how much labeling actually costs. Research from NYU Stern and Emory University shows that AI disclosures can reduce advertising effectiveness by up to 31.5%.
That is, the question of where to draw the line cannot be separated from the question of how much a marketer is prepared to lose if answered honestly.
“How are marketers reacting to this, considering there is currently a lot of concern in the market around AI and its impact on society and the economy,” said Nada Bradbury, CEO of AD-ID.
The industry is trying to work it out. Trade bodies are developing guidance, brands are taking their own policy, and regulators are lobbying to implement something more proportionate than the vague instruments currently in place. But the process is slow and the clock is not. The European Union’s AI transparency provisions reached their compliance deadline in August. New York state is requiring AI-generated human disclosures in marketing starting in June. These unfinished lines drawn by the industry will be picked up by people with completely different priorities.
“Over the last few months alone, I must have had five, six, or seven brands ask, ‘do we have guidance?’ and ‘do we have insight into how brands are thinking about this issue?’,” said Gabrielle Robitaille, policy director and AI community lead at The World Federation of Advertisers.
The reason, he says, is the same reason that drives most conversations about compliance: proximity. As August approaches, legal teams that previously left questions to marketers are now demanding answers, and marketers are looking for where to turn. Anyone who has experienced the passage of the General Data Protection Regulation will be aware of that feeling. The question now is what exactly marketers should do with the time they have left.
The WFA’s answer was intentionally narrow. The starting point, Robitaille said, is that the basic principles of advertising self-regulation apply no matter what technology is used. Don’t use it to mislead. Do not use it to make unsubstantiated claims or exaggerate product efficacy. If you remove all potentially harmful use cases, he says, the remaining gray area is actually quite small.
All that remains is a matter of threshold. The WFA is of the view that labeling should be applied when AI plays a significant role in a message of a commercial nature – that is, when AI materially shapes what someone believes about a product.
Below the line, disclosure is a decision. On it, the responsible step is to label it. The clearest case, says Robitaille, is synthetic humans, where societal anxiety occurs consistently and is well documented across studies. The beach backdrop generated behind a shampoo bottle is an entirely different matter – it’s technically AI but it’s unlikely to influence anyone’s purchasing decision.
WFA is not alone in trying to solve this problem. In January, the IAB published what they called the industry’s first unified standard for AI disclosure in advertising – and broadly speaking, they achieved the same goal. Don’t label everything. Focus on content that can actually be deceptive. Treat synthetic humans as the most obvious trigger. Background changes, audio enhancements, and post-production adjustments – all fall under the bounds of what the IAB calls “standard production techniques.”
Furthermore, this is explained specifically. Digital twins of live humans depicted in fake events should be labeled as well as AI-driven images and videos on the fly. Deceased individuals rendered by AI require disclosure, the IAB also requires C2PA metadata to be included in each ad — so that platforms like Meta, TikTok, and YouTube can see exactly what AI was used and apply their own labels if necessary.
“Advertisers’ response to this framework has been thoughtful and pragmatic,” said Caroline Giegerich, vice president of AI at IAB. Most brands don’t want to hide their use of AI and are instead looking for clarity. The real question facing brands is not ‘Should we reveal this?’ but ‘When does disclosure provide meaningful benefits to consumers versus when does it cause confusion or fatigue?’ Currently, there is an understandable sensitivity to labeling something as AI-generated because the term can carry unintended connotations, ranging from inauthenticity to manipulation.”
If the term AI has those connotations, some brands have drawn the obvious conclusion: don’t associate yourself with AI at all. Enter the “No AI” disclaimer – the certified organic label from the ad.
For example, Aerie, American Eagle’s best friend brand. They have included Pamela Anderson in an advertisement to promote the commitment they made last October not to use AI-generated bodies or humans. In it, the actor is shown encouraging the chatbot to create a model before revealing that they were real people all along. The promise is an extension of the brand’s 2014 promise not to retouch people in its ads. For Aerie, “No AI: it’s not a compliance decision, it’s the brand’s DNA.
Baby products brand Coterie has gone even further, by publicly committing to not using AI-generated images in its social media marketing at all – a stance that Coterie CEO Jess Jacobs told the Wall Street Journal is a play on trust in a crowded market where parents are a highly skeptical audience. Le Creuset has gone to great lengths to clarify that its latest social content – an inventive visual video by digital artist Ian Padgham, which sees the brand’s signature cookware transformed into unexpected objects – does not involve any AI. Clarifications appeared in the comments on the post, unpromoted, before anyone could assume otherwise.
This is a remarkable change for an industry that spent three years telling clients that AI would change everything.
“Over time, consistent transparency will normalize AI’s role in advertising, just as we have normalized other technologies before,” Giegerich said. “The goal of the industry should not be to over-label or under-label – it should be to express information in a way that provides information without being overwhelming and builds confidence, not doubt.”
Numbers you need to know
$30 billion: Anthropic’s reported run-rate revenue
$102 billion: Total advertising revenue that OpenAI estimates by 2030
61%: Percentage of US retail business decision makers who use media mix modeling to measure incrementality
78%: Percentage of US millennials (not Gen Z) who are more likely to use a second screen during 2026 World Cup games
What we have discussed
Who is OpenAI’s global head of advertising, David Dugan?
We know that Application CEO Fidji Simo has been looking for an advertising boss since last September. Now, David Dugan, has been tasked with the role – Digiday spoke to a number of people who know him to understand why OpenAI feels he is the right man for the job.
TikTok’s Khartoon Weiss will leave the platform as his North American company will lead the global business
Another top executive leaving TikTok as Khartoon Weiss will leave the platform today, according to a memo seen by Digiday, after spending nearly six years in the business.
The Trade Desk is changing how advertisers buy — and what they can see
Within the Kokai platform, The Trade Desk has tested two new automated buying modes that simplify campaign management, but combine media, data and technology costs into one price.
What OpenAI’s TBPN deal reveals about the limits of branded entertainment
TBPN’s value lies in the fact that Meta CEO Mark Zuckerberg, who rarely makes the news, has sat down with its hosts. That’s what Microsoft boss Satya Nadella has. Altman himself keeps coming back.
What are we reading
Why Anthropic’s new model has cybersecurity experts confused
Anthropic is effectively saying that the model is too dangerous to release to the general public, as a model of this caliber would soon be accessible to criminals, hackers, and countries, according to Platformer.
OpenAI asked California and Delaware to investigate Musk’s ‘anti-competitive conduct’ ahead of a trial in April
OpenAI has sent letters to the attorneys general of California and Delaware, urging them to investigate anti-competitive behavior by Elon Musk and his colleagues, according to CNBC.
Why podcast acquisition is actually the first step towards building an AI advertising stack
AI Advertising Economics outlines how OpenAI’s acquisition of TBPN supports the AI platform’s advertising ambitions.
Perplexity’s monthly revenue jumped 50% in the shift from search to AI agency
The AI platform’s estimated annual recurring revenue increased to more than $450 million last month, after it launched new agent tools, and shifted to usage-based pricing, according to the Financial Times.
PakarPBN
A Private Blog Network (PBN) is a collection of websites that are controlled by a single individual or organization and used primarily to build backlinks to a “money site” in order to influence its ranking in search engines such as Google. The core idea behind a PBN is based on the importance of backlinks in Google’s ranking algorithm. Since Google views backlinks as signals of authority and trust, some website owners attempt to artificially create these signals through a controlled network of sites.
In a typical PBN setup, the owner acquires expired or aged domains that already have existing authority, backlinks, and history. These domains are rebuilt with new content and hosted separately, often using different IP addresses, hosting providers, themes, and ownership details to make them appear unrelated. Within the content published on these sites, links are strategically placed that point to the main website the owner wants to rank higher. By doing this, the owner attempts to pass link equity (also known as “link juice”) from the PBN sites to the target website.
The purpose of a PBN is to give the impression that the target website is naturally earning links from multiple independent sources. If done effectively, this can temporarily improve keyword rankings, increase organic visibility, and drive more traffic from search results.