Urgent.News

What's breaking now, across thousands of outlets.

AI

The Pope’s AI Guy Is Worried About ‘Cartel’ Behavior Among Big Labs

Paolo Benanti tells WIRED that hysteria over whether godlike AI could destroy humanity is distracting from the need for public debate about how to govern the technology.

The Pope’s AI Guy Is Worried About ‘Cartel’ Behavior Among Big Labs

Father Paolo Benanti, a priest who advises the Catholic Church on AI, expresses concern over the "cartel" behavior among major AI labs. He argues that the focus on superintelligence distracts from the need for open debate on how to limit the power of the companies developing the technology. According to Benanti, by presenting the problem as too complex for anyone else to handle, the big labs exclude everyone from the conversation.

He believes that ethics should be debated in the public square, not confined to the boardrooms of corporations. Benanti advocates for democratic regulation, suggesting that standards should not be designed by the labs themselves but rather established through broader public discourse. The AI industry's approach to superintelligence, with its language evoking divinity, may skew the discussion and lead to misplaced reverence.

Benanti emphasizes the need for brakes and a wheel, meaning human control over the technology, to ensure it benefits society.

Written by urgent.news from Wired's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at wired.com →

More in AI

Optimal liability for offensive and defensive AI

How much liability should AI providers bear when their services enable both attack and defence? Liability can improve welfare while increasing harm.

  • Liability for AI services is complex, balancing offense and defense.
  • Overly strict liability weakens defense and boosts attacker profits.
  • High liability can lead to universal guarding in pure equilibrium.

Text-to-SQL in Practice: When to Trust AI Output and When to Gate It

AI assistants are now a normal part of SQL work. You describe what you need, get a query back in seconds and move on. The problem is not that these queries fail. Most of the time, they run perfectly.

  • 84% of developers use or plan to use AI tools for text-to-SQL tasks
  • 46% distrust AI output accuracy, 33% trust it
  • Implement risk-tiered workflow and guardrails for AI-assisted SQL queries

More from Wednesday 23 September →