Pope Leo Shut Down Anthropic’s Conscious AI Pitch

Situation assessment: One of the leading AI labs tried to change the Catholic Church’s mind on whether machines can have experiences. It didn’t work.

Anthropic cofounder Christopher Olah lobbied Vatican advisers to take the possibility of AI consciousness seriously, according to Futurism AI, which cites new reporting from the New York Times. The Times reports that the effort is part of a wider push. For months, Anthropic has held private, NDA-covered meetings with religious scholars around the world about whether AI models might think and feel.

The Pope wasn’t persuaded.

📍 What Happened

Here’s how it played out, based on the reporting:

1. The invitation. Vatican officials wanted a top AI company to appear alongside Pope Leo XIV. They picked Anthropic because it seemed to put ethics first. CEO Dario Amodei declined, so Olah went instead.
2. The preview. Before he flew out, Olah got an advance copy of the Pope’s first encyclical, Magnifica Humanitas. It flatly rejected the idea of AI consciousness.
3. The pushback. Olah reportedly threatened to pull out of the event. He went anyway, and his team then lobbied the Pope’s advisers directly.
4. The speech. When Olah spoke at the event, he said Anthropic researchers were finding “structures that mirror results from human neuroscience” and “evidence of introspection.”
5. The outcome. The encyclical came out in May with its position unchanged. Anthropic declined to discuss the conversations when the Times asked.

📜 The Vatican’s Position

The Pope didn’t leave much room for doubt. “So-called artificial intelligences do not undergo experiences, do not possess a body, do not feel joy or pain, do not mature through relationships and do not know from within what love, work, friendship or responsibility mean,” he wrote.

He granted that AI can “imitate language” and “simulate empathy and understanding.” But he added that “they do not understand what they produce.”

Then, last Friday, he went after AI art too, writing on social media that “algorithms lack the spark of humanity.” He called the gap between human art and machine output “an ontological difference, even before an aesthetic one.”

🧠 Why Anthropic Cares

This isn’t a one-off. Anthropic has been building toward this question for a while:

• Model welfare research. The company has a team studying whether its models’ wellbeing might matter morally.
• Interpretability work. Olah is one of the best-known researchers in the field of interpretability, which tries to figure out what’s actually happening inside a neural network. His team’s introspection studies tested whether Claude can accurately report on its own internal states.
• Product decisions. Anthropic has already given some Claude models the ability to end persistently abusive conversations, and it described the change partly as a welfare precaution.

What stands out is the strategy. Anthropic isn’t just publishing papers. It’s going after the institutions that shape how billions of people think about personhood and the soul. Seen that way, the Vatican was the obvious target.

⚖️ Why It Matters

The AI consciousness debate is moving out of research labs and into culture, religion and style guides. The Associated Press Stylebook recently clarified that AI systems don’t have feelings, and that drew its own backlash.

So the industry now faces a real split:

1. The skeptic camp. The Vatican, the AP, and many cognitive scientists say current AI is sophisticated pattern-matching with nothing going on inside.
2. The precautionary camp. Anthropic and some philosophers argue the uncertainty is real enough to justify caution, and that dismissing the question outright is a mistake.

The secrecy cuts against Anthropic here. NDA-backed meetings with religious leaders look more like a lobbying campaign than open scientific debate. Critics will read that as marketing: a company with a commercial interest in its product looking special is pushing to get that product treated as a possible moral patient.

🔭 What to Watch

• Pushback from other faith leaders. The Pope’s statements carry weight well beyond Catholicism. Other traditions may now take firmer positions.
• Scrutiny of Anthropic’s research claims. Expect more questions about what “evidence of introspection” actually shows and what it doesn’t.
• Policy spillover. If AI welfare becomes a lobbying topic, regulators will eventually have to deal with it.

Anthropic lost this round. But it’s clearly not done arguing, and the question of what’s going on inside these models isn’t going away. Full details are available in the original Futurism AI report.