Pangram: AI Detection Gold Standard or Flawed Arbiter?
Original: Pangram Has Emerged as the Gold Standard of AI Detection. Should You Trust It?
Why This Matters
AI text detection is becoming a gatekeeping force in publishing, with real career and financial consequences for authors.
Brooklyn-based AI startup Pangram, with 24 employees and $13M raised, has become a de facto authority in AI-generated text detection. Its percentage-based scores have led to canceled book deals and publishing scandals, prompting both adoption by Substack and fierce criticism from writers.
Pangram is a 24-person AI startup headquartered in Brooklyn that has raised $13 million total, including a $9 million round in July 2026. The company's tool claims to detect the percentage of AI involvement in any given text. Despite its small size, Pangram has become central to several high-profile publishing controversies. In January 2026, CEO Max Spero posted on X that the novel 'Shy Girl' by Mia Ballard was 78% AI-generated; publisher Hachette subsequently canceled the release. A New York Times Modern Love column scored 100%, as did a Commonwealth Short Story Prize winner. A thriller titled 'Call Me, I'll Hide the Body,' which sold for $2.4 million, scored 97%. In late July, Substack announced integration of Pangram into its platform. Spero, 30, co-founded Pangram with Bradley Emi after both attended Stanford; Spero previously worked at Google on the FLoC project. In July, the company also launched Pangram 4, its newest detection model, and posted six new job openings—a potential 25% headcount increase. Critics, including publishing expert Jane Friedman, describe widespread anger toward AI detection software, with some viewing companies like Pangram as equally problematic as AI developers themselves.