Pangram Raises $9 Million to Combat 'Dead Internet' AI Slop
The AI detection startup is partnering with Substack to help users identify synthetic content as bot-generated traffic floods the web.
Max Spero, CEO of AI detection startup Pangram, warns that the internet is dangerously close to the "dead internet theory" becoming a reality. The company is positioning itself as a critical trust layer for a web increasingly saturated by synthetic content.
To scale its efforts, Pangram has raised $9 million in funding led by Menlo Ventures, with additional participation from Haystack, ScOp, Script Capital, and Cadenza. The capital will support the development of advanced tools designed to filter "AI slop" from high-stakes environments, including insurance claims, job applications, and product reviews. As part of this rollout, Pangram launched "Pangram 4," a text detection model that the company claims achieves over 99% accuracy in identifying AI-assisted and mixed human-AI content.
The Battle for Human Signal
The company's mission is rooted in the "Dead Internet Theory," the belief that the majority of internet traffic and content is generated by bots rather than humans. This concern has intensified as Large Language Models (LLMs) enable the mass production of SEO-driven "slop" and sophisticated disinformation campaigns. Pangram was founded by Stanford AI/ML graduates Max Spero and Bradley Emi to provide a mechanism for distinguishing human intent from synthetic output.
"If we do not actively discriminate in favor of human content, then we’re just gonna see more and more AI, and it’s just gonna drown out any human signal that we have," Spero told TechCrunch.
Integration and Image Detection
In a move to bring these tools to the public, Substack has integrated Pangram's technology into its platform. The partnership allows readers to scan posts, comments, and replies—specifically those over 100 characters—to determine if they were written by AI. Substack CEO Chris Best emphasized the importance of human creativity, stating that while software should handle the mundane, the "hard part" of generating an idea worth sharing must remain with the person.
Beyond text, Pangram is tackling synthetic imagery. Unlike many detectors that rely on digital watermarks, Pangram's image tool analyzes pixel-level distributions. This approach allows the software to detect AI-generated elements even when they are embedded within otherwise real photographs.
Why It Matters
The proliferation of AI content threatens the fundamental utility of the internet as a source of human connection and knowledge. As synthetic text floods critical sectors such as journalism, academia, and law, the ability to verify human authorship becomes essential for maintaining institutional trust. If human signals are completely overwhelmed by an exponential increase in AI output, the web risks becoming a closed loop of bots communicating with other bots.
What's Next
Industry observers are now watching whether detection tools can keep pace with the rapid evolution of LLMs. While Pangram's 99% accuracy claim is a significant benchmark, the ongoing arms race between AI generators and detectors remains a primary challenge. The company's success will likely depend on whether platforms like Substack can successfully incentivize human-centric writing over automated efficiency.