Pope’s Anti-AI Warnings May Be AI-Written, Detection Tool Claims

Even the Pope’s warnings astir artificial quality whitethorn person been written by artificial intelligence.

First reported by Wired, Pangram Labs, an AI-detection startup, said its updated Chrome hold flagged respective posts from Pope Leo XIV’s authoritative X relationship arsenic apt AI-generated, including messages informing followers astir AI’s effect connected quality thought and societal structures.

One of those posts warned: “When simulation becomes the norm, it weakens the quality capableness for discernment.”

“Clearly, helium doesn't tally his Twitter relationship himself,” Pangram Labs CEO Max Spero told Wired. “They person a societal media person. But it's besides evident that they usage astatine slightest immoderate grade of AI.”

Pangram claims its detection exemplary has a 99.98% accuracy complaint and a mendacious affirmative complaint of 1 successful 10,000. A December 2025 University of Chicago study ranked Pangram highest among tested AI detection tools and recovered its mendacious affirmative complaint was adjacent to zero.

“Given that each detector is astatine slightest somewhat imperfect, organizations inactive person to measure for themselves if and however to usage them, trading disconnected the imaginable for AI misuse with the hazard of mendacious accusations,” the survey said.

Pangram Labs did not instantly respond to Decrypt’s petition for comment, nor did representatives for the Vatican.

AI detection remains controversial, and mendacious positives stay a interest crossed the industry. In 2024, AI detector ZeroGPT made waves online aft it was revealed by Christopher Penn, Chief Data Scientist astatine Boston-based selling analytics steadfast Trust Insight, to person labeled the U.S. Declaration of Independence 97.93% AI-generated.

“These tools are being utilized to bash things similar disqualify students, putting them connected world probation oregon suspension,” Penn told Decrypt astatine the time. That’s “a precise high-risk exertion when, successful the United States, a assemblage acquisition is tens of thousands of dollars a year,” helium said.

According to Wired, Pangram’s hold besides reportedly flagged posts from blue-check influencers connected X, trending Substack writers, and a connection from Apple CEO Tim Cook marking the company’s 50th anniversary.

The motorboat comes arsenic concerns implicit AI-generated junk contented proceed to rise. Earlier this month, researchers astatine Stanford University, Imperial College London, and the Internet Archive found that AI-generated oregon AI-assisted substance accounted for astir 35% of recently published websites by mid-2025.