Artificial intelligence might have contributed to the Pope’s own warnings about its influence, according to a detection startup.
As reported initially in Wired, Pangram Labs, an AI-detection company, indicated that its updated Chrome extension identified several posts from Pope Leo XIV’s official X account as potentially being AI-generated. These included messages cautioning followers on how AI could impact human cognition and societal frameworks.
One such post cautioned: ‘When simulation becomes the norm, it weakens the human capacity for discernment.’
‘Clearly, he doesn’t personally manage his Twitter,’ Max Spero, CEO of Pangram Labs, remarked to Wired. ‘They have a social media representative, but it’s clear they employ at least some AI tools.’
Pangram boasts a 99.98% accuracy rate for its detection model, with a false positive rate of one in 10,000. A study from the University of Chicago in December 2025 ranked Pangram as the top AI detection tool among those tested, noting that its false positive rate was nearly zero.
‘Though every detector has minor imperfections, organizations must decide on their use based on balancing the risks of AI misuse against potential false accusations,’ the study noted.
Neither Pangram Labs nor Vatican representatives responded to Decrypt’s request for comments.
AI detection continues to spark debate due to concerns about false positives. In 2024, ZeroGPT, an AI detector, garnered attention after being exposed by Christopher Penn of Trust Insight as having identified the U.S. Declaration of Independence as 97.93% AI-generated.
Penn highlighted the risks involved in using these tools for high-stakes situations like academic evaluations: ‘Disqualifying students and placing them on probation or suspension is a serious application when college education costs are substantial,’ he told Decrypt.
Wired also reported that Pangram’s tool flagged posts from verified X influencers, popular Substack writers, and an anniversary message by Apple CEO Tim Cook for the company’s 50th year.
The rise in AI-generated content remains a concern. Recent research by Stanford University, Imperial College London, and the Internet Archive found that nearly 35% of new websites published mid-2025 were created or aided by AI.