We’re ‘dangerously close’ to dead internet theory, says Pangram’s CEO

The "dead internet theory" - a fringe idea that once seemed like paranoia - holds that most online content is generated by bots and automated systems rather than real people. According to Pangram Labs CEO, that scenario is no longer hypothetical. In a recent conversation, she warned that the volume and quality of AI-generated content has reached a point where the line between human and machine participation online is becoming genuinely difficult to draw.

What makes this moment distinct from earlier waves of spam or bot activity is the breadth of contexts being affected. AI-generated text and images are now appearing in places where authenticity carries real consequences - hiring pipelines, consumer review platforms, and insurance claim submissions among them. These are systems that were built on the assumption that a human being stands behind each submission, and that assumption is under increasing pressure.

Pangram Labs is among a small group of startups attempting to build reliable detection infrastructure for AI-generated content. The technical challenge is substantial: generative models are improving quickly, and detection tools must keep pace with outputs that are becoming harder to distinguish from human writing or photography. The company focuses on identifying statistical and stylistic signals left behind by large language models and image generators, though no detection method is considered foolproof at this stage.

The broader concern raised by Pangram's CEO is less about any single platform and more about aggregate trust in online information. When users cannot reliably determine whether a review, a resume, or a claim was written by a person or produced by a model, the credibility of the entire system erodes. Whether detection technology, platform policy, or some combination of both can reverse that erosion remains an open question - but the conversation is no longer happening at the margins.