The AI 'nightmare' on the Meta ecosystem

Mark Zuckerberg reveals his ambition to transform the advertising industry with AI. Photo: Passionfru.it .

Speaking on stage at Stripe's annual Sessions conference in San Francisco on May 6th, Zuckerberg presented a plan to automate the entire advertising industry. The entire process, from start to finish, would be handled by a "black box" AI.

A key component of this product is the delivery of thousands of "test" ads to Facebook, Instagram, and Threads users. AI handles everything from idea generation and content writing to image/ video design, testing, and performance optimization, until the ads are distributed to the required number.

According to Zuckerberg, the goal is to satisfy all businesses, whatever their needs, and whatever amount they are willing to spend to achieve those results. “In a way, this is like a business results machine. I think it’s going to be one of the most important and valuable systems ever built,” he said.

If his vision on Ben Thompson's Stratechery podcast last week is correct, this hypothetical "machine" will have a huge impact on the advertising industry. Creative advertising agencies will still exist. However, their role will be greatly diminished, at least in terms of content creation.

At the Sessions event, Zuckerberg asserted that many of Meta's advertising tools have integrated AI-generated targeting capabilities. This functionality is so sophisticated that the company no longer even encourages customers to manually specify their target demographics.

According to him, Meta's tools can identify interested users more effectively than human advertising staff. The logical next step is to apply this data-driven optimization capability to content creation. "We can create around 4,000 different versions of a creative ad, then test and determine which one works best," Zuckerberg said.

This solution might be attractive to businesses. However, the problem with AI-powered advertising lies in the user experience on Meta's platforms, especially given the already overwhelming amount of AI-generated spam content.

Meta AI's blue-purple circle is densely packed across platforms.

Meta AI was launched in October 2024. By early 2025, users had begun to notice its frequent appearance. From posts and comments to private messages, the blue-purple circles were everywhere. Polly Hudson, a reporter for The Guardian, said she was obsessed with them and spearheaded a movement to boycott Meta AI.

Recently, the company also launched a social network dedicated to AI content, where users can share their ideas and commands. The Verge commented that the platform resembles a poster summarizing all the complaints about AI, with repetitive, boring, or fault-finding content.

According to Techcrunch , this demonstrates that Meta is a marketplace, users are the products, and customers are the advertisers. Now, this ecosystem is about to add another layer of AI-generated spam content.

The advertising industry also shows many conflicting views regarding Zuckerberg's ambitions on ethical issues. In October 2024, more than 11,000 creators signed an open letter condemning the use of human-created work to train AI systems.

On the other hand, many creators and advertising executives believe that AI tools threaten their livelihoods. Johnny Hornby, founder of the advertising agency The&Partnership, published an editorial on May 6th arguing that building successful brand campaigns remains a uniquely human endeavor.