Pope's AI Advisor Says Big Labs Are Behaving Like a Cartel

Father Paolo Benanti, an adviser to the Catholic Church on artificial intelligence matters, is sounding an alarm about how the world's most powerful AI companies are shaping the future of the technology—and who gets a seat at the table.

Rather than fixate on speculative dangers like superintelligence, Benanti argues that the real issue is far more immediate: a handful of corporate labs are setting industry standards without meaningful input from the broader public, effectively operating as what he describes as a cartel.

The Problem With Self-Regulation#

Benanti contends that when major AI firms like OpenAI and Anthropic present their safety concerns and development guidelines, they're simultaneously claiming exclusive authority to solve these problems. "Ethics cannot be something that belongs to the board of trustees of a company," Benanti told WIRED. "It's something that has to be balanced, mitigated in debate in the public square."

The priest argues this dynamic strips ordinary citizens, policymakers, and institutions of meaningful influence over how AI develops, leaving such decisions to executives and engineers who have obvious financial incentives in the outcome. He's skeptical of proposals like mandatory third-party model evaluations or restrictions on certain AI-training techniques, viewing them as industry-friendly compromises rather than genuine democratic control. What matters instead, he says, is ensuring that humans retain true oversight and decision-making power.

Profit Incentives Cloud the Narrative#

Benanti points out that the repeated warnings from AI labs about existential risks arrive at a convenient moment: while both OpenAI and Anthropic prepare for public offerings worth billions of dollars. The apocalyptic framing around superintelligence, he suggests, may distract regulators and the public from scrutinizing these companies' market dominance.

Complicating the picture further is geopolitical rivalry. As the US and China race to develop advanced AI models, national security anxieties amplify the sense of urgency—and the willingness to let private companies lead the charge.

A Different Framing#

Benanti also challenges the reverent language some industry leaders use when discussing AI capabilities. When CEOs speak of building "magic intelligence in the sky" or treat artificial systems with quasi-religious awe, it obscures what AI actually is: sophisticated pattern-matching, not divinity. "We are not building the next divinity," he emphasizes. "We are simply deluding ourselves that there is something more than statistical computation [going on]."