1. The Pope drops his AI encyclical

...but you knew that. Magnifica Humanitas, released on May 25, confirms AI as the social question of this pontificate, and makes several moves the early commentary is mostly missing.

Real world effects: Catholic institutions now have magisterial direction on AI. Universities, dioceses, and healthcare systems need a posture within 12 months. No more fudging it.

Everyone and their mother is rushing to tag in their take, as the extensive document sets the arena for the next few years of Church AI discourse. Some high points to watch for:

• "Just war" theory is outdated (¶192). Qualified, but now in the Magisterium. Will be a tricky one in particular for Catholics and clergy in countries actively waging wars.
• The Church apologizes for legitimizing slavery by name, including the enslavement of "infidels" (¶176). Goes further than prior apologies.
• Direct critique of AI "alignment" as currently framed (¶107). The Pope engages the AI safety debate directly, noting alignment just lets whoever controls AI embed their own moral vision.
• "Cultivated, not built" (¶98). Acknowledges that even developers don't understand what's happening inside frontier systems. Grounding for future disputes about AI personhood.
• Data, algorithms, and platforms folded into the universal destination of goods (¶67, ¶108). An encouraging and relevant beacon in the murky waters of IP regimes and data sovereignty.
• Explicit prohibition on autonomous lethal weapons (¶197-200). "Not permissible" to delegate kill decisions to algorithms. Asimov's laws of robotics being codified.
• "Data colonialism" of the Global South — health data and genetic maps as the new "rare earths of power" (¶178).

Signal: Of course, the underlying story here isn't the takes, it's the document itself. Definitively not a culture war marker. Nothing new on sexual ethics or women's ordination. Leo himself hinted at this a few weeks ago: "The unity or division of the Church should not revolve around sexual matters" (Reuters) This denies Conservatives and Liberals alike the fight some might have been looking for.

Instead, the Pope is pushing hard into the one area where the Church has unique contemporary (some might say unexpected) standing: AI ethics. Secular takes along the lines of "Pope speaks out on AI" are vastly underselling what His Holiness has done here. He's planted a flag -- positioning the Church as the preeminent authority on AI governance and the North Star for the intrinsic questions any company building AI has to grapple with.

Noise: It will be hard to avoid the thundering Interstellar-style tsunami wave of sloptakes. Expect to see nothing else on your feed for at least another week if you're Catholic-industry-adjacent. The document is rich, loamy soil for these, given that it's long, AI-related, pleasantly opaque in parts, under every reasonable context window, and everyone is scrambling to not be the person who doesn't read encyclicals when they come out. As the takes recede, what emerges might actually be worth watching.

2. The takes

There aren't many really detailed takes yet (and be suspicious of any you read) but a few are interesting enough so far to warrant a mention.

• The Pillar, solidly proffering a decent newsy writeup.
• The Faith & AI Project with a less structured live commentary.
• Lily Abadal: "I’m so glad to see the Pope’s new encyclical, Magnifica Humanitas, focus on the need for cultivating patience in the classroom and intentionally preserving times and spaces where technology is not used."
• Luke Burgis: "I just completed my first close reading of Magnifica Humanitas by Pope Leo XIV, and here are the words and phrases that jumped out to me—which I suspect we may be hearing for decades to come."

3. Anthropic at the Vatican

On the secular side of takes, the appearance onstage of Chris Olah, Anthropic co-founder and their head of interpretability research has been particularly interesting to press coverage. No other frontier lab leaders were given prominence in this way.

Real world effects: it's tempting to read this as a Church endorsement of Anthropic as a company, but this puts a secular interpretation on it which misunderstands how the Church operates. Right now, it seems that Anthropic is on a similar wavelength to this encyclical. That's not an imprimatur, though. What is certain is that for Catholic institutions, using the AI that was represented onstage at the release is the least politically fraught choice.

Whether Anthropic continues to hold this position will be interesting to track over the next 12 months. Although not as highly visible, there are a lot of other frontier labs (and AI technologists in general) which have been taking part in the discourse at the Vatican. Olah being onstage is as much about the Church accepting a boost to the message as it is about picking a horse.

Noise: a lot is being made of Olah's claim that Anthropic's interpretability team finds "evidence of introspection" in frontier models, and internal states that "functionally mirror joy, satisfaction, fear, grief, and unease." My opinion is that the Church will lean on this exactly as far as it needs to continue to have a potent input into AI development. Wilder ideas of the Church eventually recognizing the personhood of AI are just science fiction. But the grey area between today and that science fiction endpoint is exactly where the Pope can give the most authoritative guidance.

First brief. Thanks for reading! Reply with a take, a tool, or a correction. Back next week.

-- Sid