AI Advertising: A Major Marketing and Security Challenge

The Emergence of Ads for AI Agents

A new type of advertising is emerging, targeting not human consumers but artificial intelligence agents. This phenomenon opens a new front in the marketing field while posing security challenges for brands. The crucial question is how AI platforms will respond to this practice: will they accept this new channel or seek to eliminate it?

A Concrete Example of AI Advertising

To illustrate this concept, the Adtech startup Mobian recently collaborated with Time to implement "Agent Ads" for brands such as Ally Bank and the Project Management Institute. Time created simplified versions of its web pages, specifically designed to be read by AI agents, devoid of graphical elements that humans would normally see. These pages, invisible to human readers, are now used to deliver advertisements in the form of text blocks optimized for AI bots.

Influence on Chatbots and Brands

These ads could allow companies to leverage the authority of publishers like Time to influence the perception of their brands in chatbots such as ChatGPT and Gemini. Andrea Brimmer, Chief Marketing Officer of Ally Bank, emphasized the importance of these signals, noting that visitors from AI platforms were 3.5 times more likely to open an account than those from traditional searches. Furthermore, traffic generated by AI has seen a ninefold increase year-over-year.

A Heated Debate in the Industry

The introduction of ads for AI agents has sparked an intense debate within the industry, particularly on social media and in private discussions. Questions raised include the legitimacy of this new advertising channel and the risk of malicious actors creating ad networks to promote scams or manipulate AI recommendations to the detriment of competitors. Will the industry be able to react quickly to establish standards before AI companies block these practices as spam?

Reputational Risks and Historical Precedents

Brands that attempt to discreetly manipulate the situation expose themselves to reputational risks. The example of Bell Pottinger, a public relations agency that secretly altered Wikipedia pages for its clients, remains a cautionary tale. Bernard Urban from BSCI warned of potentially swift and severe public backlash against such practices.

Transparency and Accountability

Menaka Gopinath, Chief Marketing Officer of PMI, highlighted that AI, like many online media, raises questions of transparency and accuracy. She appreciated the approach taken by Time and Mobian, which ensured that sponsored content was clearly identified. According to her, organizations, publishers, and technology providers all have a role to play in establishing trust.

Policies of AI Platforms

The most popular AI platforms have claimed that ads do not influence their responses, although this seems to apply to ads displayed within their own chatbots rather than those on websites targeting their crawlers.

The Influence of Advertising on AI Models

Jonah Goodhart, CEO of Mobian, stated that AI models are clearly influenced by advertising, as they are trained on what people write and publish. Advertising thus shapes culture and public record. Mobian offers "clearly labeled facts, brand-verified, supported by sources," whose impact depends on the context and credibility of the sources.

Accountability and Brand Safety

Andrea Brimmer emphasized that every new technology brings new responsibilities. She insisted that ads aimed at LLM bots undergo the same legal and compliance checks as those aimed at consumers. According to her, ignoring what LLMs say daily about a company presents a greater brand safety risk.