AI essentials: agents, AI influencers, regulation targets use of synthetic humans, Meta EU rollout, in-house agencies

Are AI Agents clicking on your ads?

A recent study has revealed that AI agents like OpenAI’s Operator, Anthropic’s Claude and Google’s Gemini 2.0 Flash - which are capable of autonomously performing tasks and making decisions on users’ behalf - are not only browsing the web; they’re actively clicking on ads and making purchasing decisions.

In a simulated hotel booking scenario, these agents completed tasks such as ‘book me the cheapest romantic holiday’. The study found that while there were significant differences in how each of the Agents engaged with ads (and differing conversion rates), keyword matching significantly impacted advertisement effectiveness.

This development highlights a shift from traditional ‘User Experience’ to ‘AI Experience’ and as AI agents become more prevalent, brands must consider how their digital touchpoints are perceived by both humans and machines.

WFA will be discussing AI agents at the next AI Community meeting, taking place on 11 June. You can register here. And if you’re in Cannes, we’ll be co-hosting an exclusive event with Meta on the opportunities of Agentic AI and what it means for marketers. If you’d like to participate, reach out via email.

Brands raise concerns about AI influencers

WFA’s latest research finds that AI influencers have yet to win over most major brands. Just 15% have experimented with them so far and 60% say they have no plans to use them. The research, which explores marketers’ views on the opportunities and challenges of AI influencers, highlights concerns regarding the potential implications of using AI influencers (including existing ones such as Miquela, bespoke AI-created brand ambassadors or AI-powered livestream hosts).

While a majority recognise the main opportunities as cost efficiencies (80%), protection from influencer scandals (58%) and scalability (58%), 96% cite concerns around consumer trust and acceptance, with issues around authenticity (73%) and potential brand reputation risks (58%) also weighing heavily on marketers’ minds.

The full findings are based on responses from 27 of the world’s biggest companies.

New York Senator introduces draft bill on the use of synthetic humans in ads

A draft bill has been introduced in the New York State Senate that would require marketers to clearly disclose the use of ‘synthetic performers’ in advertising. The proposed law, introduced by Democrat Senator Michael Gianaris, defines a synthetic performer as ‘digital asset that is created, reproduced, or modified by a computer using generative artificial intelligence […] that is intended to give the impression that the asset is of a human performer’. Failure to disclose would result in a $1,000 fine for a first violation and $5,000 for any subsequent offense.

This draft legislation reflects growing regulatory interest in the use of AI in marketing and understanding when and when not to label and disclose AI-generated marketing content is a growing priority for brands. To support members in maintaining trust and transparency in their use of Gen AI, WFA’s AI Community is developing best practice guidance on disclosure.

Meta begins AI rollout in the EU

Meta announced that it will start training its AI models on the data of EU users of Facebook and Instagram. Meta says it will include data from public posts, comments and chat history with its AI chatbot, Meta AI, but will not include private messages. EU users will be notified via in-app notifications and email and will be provided the opportunity to opt out via an objection form.

Meta put its AI training plans in Europe on hold last year after its EU privacy regulator, the Irish Data Protection Commissioner (DPC), raised concerns that its training methods were not compliant with existing data protection and privacy laws.

According to Meta, this latest decision follows ‘constructive’ engagement with the Irish DPC and an opinion provided by the European Data Protection Board, bringing together all EU privacy regulators.

How are in-house agencies using Gen AI?

In response to growing interest from WFA members, we’ve launched a global survey in partnership with R3 focused on how in-house agencies are using Generative AI today.

In-house agency experts can respond to the survey here, to help shape the industry’s view on Gen AI and benchmark the practices of their team against other leading in-house agencies.

In particular, the survey explores how in-house teams are applying Gen AI, what tools are most commonly used, what tasks Gen AI is best suited to take on and where budgets are coming from. The survey should take no more than seven minutes and all answers will remain confidential.

WFA’s next In-House Agency forum will be taking place on 8 May, where we’ll be investigating how brands can best future-proof their own internal marketing ecosystem in the age of AI. You can register here.

Please send across any tips, developments and interesting insights to Gabrielle Robitaille.