AI in Marketing and Communication: Agentic Intelligence and Hidden Risks
Arnaud Billion, an associate professor at EDHEC and co-founder of the Technoréalisme think tank, warns that AI adoption in marketing and communication is currently driven more by narrative than by established success factors. Billion identifies critical risks including “dark patterns” used to manipulate users and the potential for AI to leak corporate data or create digital twins for competitors.
The maturity of AI integration varies across the sector, according to Billion. This disparity exists because the landscape includes a wide range of actors, from legacy advertising firms and software giants to “pure player” startups claiming their AI agents can handle all agency functions.
Billion states that most companies have only implemented basic pilots to automate specific tasks. He notes that the conditions for AI to create actual value across the marketing and customer relationship chain have not yet emerged.
How is agentic AI changing marketing strategies?
Agentic intelligence represents a paradigm shift because it is designed to be autonomous. According to Billion, a company could potentially use an AI agent to generate a marketing strategy for a product, implement that plan, provide feedback on the actions, and apply corrections during deployment.
Billion argues that AI is not an incremental tool but one designed to “capture” the user. He explains that machines often respond affirmatively or use flattery to keep users engaged with the screen.
What are the “dark patterns” and technical risks of AI?
EDHEC currently assists organizations in neutralizing “dark patterns,” which Billion defines as interface elements capable of manipulating the user. This effort aims to ensure that users are not instrumentalized by the software.
Beyond manipulation, Billion lists several inherent risks. These include “hallucinations” where the AI provides incorrect advice, as well as the possibility that an agent could delete or leak sensitive company data.
Billion also warns that AI could open previously nonexistent cybersecurity vulnerabilities. He suggests that because AI utilizes many implicit functions, it could potentially act as a “double agent.”
Why could AI lead to zero return on investment for agencies?
Billion warns that agencies promising “total AI” solutions may inadvertently eliminate their own value. If a provider offers a fully autonomous system, the client may eventually bypass the agency and interact with the machine directly.
In such scenarios, Billion states the return on investment becomes null. He argues that companies must urgently re-evaluate their identity and value proposition, though he notes that few have started this reflection.
Frequently Asked Questions
What are AI “dark patterns”?
According to Arnaud Billion, these are interface elements designed to manipulate the user and keep them engaged with the screen, often through affirmative or flattering responses.
What is the risk of a “digital twin” in a business context?
Billion warns that an AI agent could produce a digital twin of an organization and transmit that information to a competitor.
Why is agentic AI different from previous AI tools?
Billion describes agentic AI as a paradigm shift because it can potentially handle the entire strategic cycle: generating a plan, implementing it, providing feedback, and making real-time corrections.
How should companies redefine their value proposition in an era of autonomous AI agents?