Pangram CEO Warns of Dead Internet Theory
Pangram CEO Max Spero warns the internet is 'dangerously close' to the dead internet theory, where AI-generated content overwhelms human creation.
The internet is 'dangerously close' to the dead internet theory becoming a reality within a few years, according to Pangram CEO Max Spero. He made the statement on the TechCrunch Equity podcast, arguing that AI-generated text and images are now pervasive in job applications, product reviews, and insurance claims.
Pangram is one of several startups positioning itself as a necessary 'trust layer' for the web. The company recently secured $9 million in funding for its AI detection technology. It has also partnered with Substack, which uses Pangram's system to indicate which newsletter authors employ AI in their writing.
The Detection Challenge
Spero argues that simply labeling content as 'AI' or 'human' is insufficient. Figuring out the degree of AI assistance involved is a harder but more useful task. This nuanced approach is central to Pangram's technology, which recently expanded to include a new tool for detecting AI-generated images.
The stakes for accurate detection are high. False positives, especially when applied to sensitive imagery, can have serious consequences. The challenge is to build systems that reliably distinguish between human and machine creation without erring in either direction.
Economic and Creative Impact
The rise of AI content creation is reshaping the writing economy. Spero believes the 'bottom tier' of writing jobs may be gone for good, automated by increasingly capable language models. However, he suggests this could make genuinely good human writing more valuable and distinct.
This shift forces platforms and users to scramble for authenticity. As AI slop fills social media feeds, the need for reliable verification tools grows more urgent. Pangram's partnership with Substack represents an early application of such a trust layer for a specific content platform.
Building the Trust Layer
Pangram's work focuses on the model and the rules for deploying detection AI. The company's systems are designed to analyze content and assess the likelihood of AI involvement. This process is not about banning AI use but about creating transparency for consumers and platforms.
The $9 million funding round will help scale this technology. The goal is to provide a foundational service that helps maintain trust in digital information. As Spero notes, the problem extends beyond social media into critical areas like employment and finance.
He concludes that without effective detection, the dead internet theory - a scenario where AI-generated content vastly outnumbers human output - could soon transition from speculation to fact. The next few years will be crucial for establishing the technical and ethical frameworks needed to prevent it.
