Why Big Tech Is Drowning in the AI Slop of Its Own Making

Why Big Tech Is Drowning in the AI-Slop It Creates Itself

In the closing days of 2025, Instagram's head Adam Mosseri published a post that can hardly be called anything other than a manifesto of digital anxiety. "Authenticity has become infinitely reproducible," he wrote. Everything that made content creators meaningful — the ability to be genuine, to establish a connection with their audience, to have a voice that cannot be faked — is now available to anyone with the right tools.

Mosseri proposed an elegant solution: camera manufacturers would cryptographically sign photographs at the moment of capture, creating a chain of trust. The result would be a reliable system for determining what is not a product of artificial intelligence. Sounds like a plan.

The problem is, it's sounded like a plan for several years running.

The Verge drew attention to the fundamental contradiction that permeates the entire industry. Companies that created and continue to scale generative AI tools simultaneously play the role of fighters against the consequences of their own technology. Meta, which owns Instagram and Facebook, actively develops its own generative models, embeds AI functions into every product, and encourages the creation of synthetic content. And the very same executive of the very same company publicly laments that authenticity is dying. This is not hypocrisy in the classical sense — it is a structural conflict of interest built into the business model.

Progress in the field of deepfake and AI-generated content labeling is indeed frustratingly slow. The C2PA standard that Mosseri references presumes embedding image origin metadata directly into the file. Several camera manufacturers — Sony, Leica, Nikon — are already experimenting with this technology. But there is a chasm between a laboratory prototype and mass deployment. The overwhelming majority of photographs in the world are taken on smartphones, and neither Apple nor Samsung have yet integrated C2PA into their devices in a way that could be called an industry standard. Moreover, metadata is easily stripped when uploading to most platforms — including, ironically, Instagram itself.

There is a deeper problem. Even if the marking of "authentic" photographs works perfectly, it does not solve the question of content that already exists. Billions of images uploaded before the standard's implementation will remain unsigned. The absence of a cryptographic signature will not mean that a photograph is fake — it will mean only that it was taken before the era of marking or on a device without standard support. A system that cannot distinguish old genuine photos from new fakes has limited value.

Meanwhile, the scale of the AI-slop problem — low-quality synthetic content filling feeds and search results — continues to grow exponentially. Facebook has already become a platform where AI-generated images get millions of likes and shares. Google's search results are clogged with AI-text sites. Amazon struggles with a flood of books written by language models. Each of these companies is simultaneously both a victim and a source of the problem — they provide tools for generating content, platforms for distributing it, and algorithms that reward its engaging properties.

Critics rightly point out that technology giants have no economic incentive to truly solve this problem. Synthetic content generates engagement, engagement generates advertising revenue, and advertising revenue is what these platforms exist for. Marking AI content, if implemented honestly and universally, could reduce interaction with such content, and therefore hurt the metrics shown to investors in quarterly reports. In this light, Mosseri's statements look not like a plan of action, but as a PR exercise: identify the problem, propose a distant technological solution, and continue to profit from the status quo.

The situation is reminiscent of how tobacco companies spent decades funding research on the harms of smoking while simultaneously continuing to sell cigarettes. The difference is that AI-slop does not kill literally — it destroys the information environment, undermines trust in visual evidence, and creates a world in which distinguishing truth from fiction becomes a task requiring technical expertise. If the largest technology companies truly want to fight this phenomenon, they should start not with cryptographic signatures on cameras, but with their own products — with limiting generative tools, with transparent marking on their own platforms, with algorithms that will not push synthetic content into the feeds of billions of users.

Until this happens, all talk of fighting AI-slop will remain exactly what it is — slop of another kind.