AI detection company Pangram has raised $9 million and secured a partnership with Substack, as its co-founder and chief executive Max Spero warned that distinguishing between human and machine-produced work is becoming increasingly difficult.
Substack is using Pangram’s technology to identify the use of artificial intelligence in newsletters, giving readers an indication of whether content from authors they follow has been written with AI.
The deal comes as AI-generated text and images spread beyond social media, appearing in job applications, product reviews and insurance claims. Platforms are under growing pressure to establish whether material submitted by users is genuine.
Pangram is positioning its software as a “trust layer” for online services. The company has also launched a tool designed to detect AI-generated images, extending its work beyond written content.
Speaking on TechCrunch’s Equity podcast, Spero discussed the limits of automated detection and the difficulty of setting a clear boundary between work created by AI and material that has merely been assisted by it.
That distinction is becoming particularly important as people use generative tools for editing, drafting and research while continuing to present the final work as their own. A system that treats any use of AI as equivalent to fully generated content risks producing misleading results.
Pangram’s technology is intended to help organisations and readers assess the provenance of digital material, although no detection system can remove the wider challenge of deciding how much reliance should be placed on such assessments.
