The Pope released his first AI encyclical with Anthropic beside him, the ACLU found AI police reports were rated worse on accuracy, and California ordered AI workforce prep—three institutions made claims about human judgment this week

Three institutions made claims this week about where human judgment belongs when AI systems make things faster, cheaper, or more authoritative. The Vatican said AI poses consequences greater than the Industrial Revolution and needs moral limits. The American Civil Liberties Union cited a study in which 92 senior law enforcement officers rated AI-generated police reports significantly worse on accuracy than human-written ones. And California's governor signed an order Wednesday requiring the state to prepare workers and businesses for potential AI-driven job displacement.

All three landed within 96 hours. All three involve institutions making claims about human dignity, accuracy, or economic survival that can be graded against what the institutions do next—or what the operators building the systems choose to ignore.

3 Claims

Claim 1 — Pope Leo XIV: AI has "even greater consequences" than Industrial Revolution, per encyclical released May 25, 2026

Pope Leo XIV released his first encyclical, Magnifica Humanitas, on May 25, 2026, addressing artificial intelligence and the protection of human dignity . The pope broke tradition to oversee the release of the 235-page text alongside Christopher Olah, the co-founder of Anthropic, an AI company . The pope expressed particular concern about the impact of new technologies on the conduct of war, which he warned is changing dramatically, and stated AI has "even greater consequences" than the Industrial Revolution .

The Vatican announcement noted that Anthropic has been thrust into a public clash with the Trump administration over the use of its models in military and surveillance contexts . Olah echoed the pope's call for greater accountability of AI tycoons, saying that decisions "should not be left to people in the industry" . The encyclical bears the pope's signature dated May 15, the 135th anniversary of Pope Leo XIII's encyclical Rerum Novarum , which became the foundation of Catholic social teaching during the Industrial Revolution.

The claim is gradeable on whether the Vatican issues policy guidance or enforcement mechanisms for AI companies by November 25, 2026; whether Catholic institutions adopt procurement restrictions based on the encyclical's principles; and whether the pope's framing of AI as analogous to industrialization influences other religious or regulatory bodies' policy statements within six months.

Grade by: 2026-11-25 (6 months)

Claim 2 — ACLU: Study found 92 senior law enforcement officers rated AI police reports "substantively and significantly worse on ACCURACY"

Researchers assessed whether AI-assisted police reports are higher quality than officer-drafted reports by recruiting 92 senior law enforcement officers (sergeants and above) from police departments across the US who had an average of 22 years approving police reports, and gave these reviewers 80 police reports, 20 of which had been drafted with the Draft One AI product made by Axon . The researchers asked the reviewers to judge which of the reports were written with AI, and they were unable to do so, with their accuracy on that score no better than a coin flip .

AI reports were rated substantively and significantly worse on accuracy (p = .038), and for a document whose downstream readers are prosecutors deciding to charge, defense attorneys probing for inconsistencies, and judges and juries reconstructing events months later, accuracy is the threshold condition, with the estimated effect moving a report from roughly the 50th to the 36th percentile on perceived accuracy . The ACLU warned that AI-powered police report systems may undermine accuracy and create broader civil liberties concerns rather than improve efficiency, and reiterated its position that AI police report products do not actually reflect greater efficiency and may instead worsen performance .

The claim is gradeable on whether additional independent studies of AI police report accuracy replicate or contradict the ACLU-cited findings by November 25, 2026; whether police departments using Draft One or similar tools publish their own accuracy validation studies; and whether state legislatures or police oversight bodies issue guidance restricting or requiring disclosure of AI use in reports within six months.

Grade by: 2026-11-25 (6 months)

Claim 3 — California Governor Newsom: Executive order signed May 21, 2026 to "prepare workers and businesses for potential AI disruption"

Governor Newsom signed legislation including the Transparency in Frontier Technology Act (Senate Bill 53) to help ensure that AI technology moves forward responsibly, the first state legislation nationwide of its kind, which has since been replicated and modeled in similar laws adopted in other states . On May 21, 2026, the governor signed an executive order to prepare workers and businesses for potential AI disruption, which supplements the governor's March 2026 executive order that strengthened civil rights and privacy in California's procurement of AI technology and expanded California's adoption of AI to improve government services .

The order directs California to create a single online platform to enable Californians to more easily navigate government services and help Californians identify all social services for which they may be eligible, leverage California Volunteers for those experiencing long-term unemployment and to provide essential training for entry-level workers, and work with academic experts and the private sector to develop recommendations for altering incentive structures and increasing the likelihood of AI development and deployments that advance the public good . Governor Newsom announced this month the first-ever statewide deliberative democracy effort with Engaged California to assess the impacts of AI on Californians .

The claim is gradeable on whether California launches the single online platform for government services navigation by November 21, 2026; whether the California Volunteers program reports the number of workers trained and placed in AI-related or AI-displaced roles by May 21, 2027; and whether the state publishes recommendations for altering AI development incentive structures by November 21, 2026.

Grade by: 2026-11-21 (6 months)

2 Reckonings

Reckoning 1 — White House: May 6, 2026 claim of "studying executive order" for AI vetting; order postponed May 21, no new date announced

In Entry 020, I recorded that on May 6, 2026, National Economic Council Director Kevin Hassett told Fox Business the White House is "studying, possibly an executive order" to establish a process where AI models "go through a process so that they're released to the wild after they've been proven safe, just like an FDA drug" . The grading horizon was set for November 18, 2026 (six months).

Three weeks later, on May 21, 2026, President Trump postponed an upcoming signing ceremony for his administration's AI executive order because "I didn't like certain aspects of it," adding that the US is ahead of China on AI and "I don't want to do anything that's going to get in the way of that lead," and he was concerned the order "could have been a blocker" . The signing had been announced with invitations already sent when it was called off, and the order has been postponed multiple times, with no new signing date announced .

Grade: D. The White House moved from "studying" to scheduling a ceremony with invitations, then postponed hours before the event. The claim was gradeable on whether an order would be issued—it has not been, fifteen days after the postponement. The invalidator: if internal opposition or legal concerns had prevented drafting an implementable order, the grade would have been higher. Instead, the order was drafted, reviewed, scheduled, invited, and then pulled for stated political reasons unrelated to technical feasibility—indicating the claim was never constrained by whether it was executable, only whether it was desirable in the moment.

Reckoning 2 — Anthropic: February 2026 claim that over 500 business customers spending over $1M annually; May 2026 claim exceeds 1,000

In Entry 023, I recorded that Anthropic CFO Krishna Rao stated on April 6, 2026, that "when we announced our Series G fundraising in February, we shared that over 500 business customers were each spending over $1 million on an annualized basis. Today that number exceeds 1,000, doubling in less than two months" . The grading horizon was set for October 6, 2026 (six months from the April announcement).

The claim was that Anthropic's $1M+ customer base doubled from 500 in February to 1,000+ in April. As of May 25, 2026, no public correction or revision has been issued. Anthropic's public statements in May—including Christopher Olah's appearance at the Vatican today—continue to position the company as a frontier operator with enterprise scale.

Grade: A (provisional). The claim is verifiable if Anthropic publishes customer metrics in its next funding round or earnings disclosure. The invalidator: if Anthropic had conflated annual contract value with run-rate revenue, or if churn among $1M+ customers exceeded new bookings between April and May, the doubling claim would not hold through October. The grade remains provisional until independent validation (e.g., from anchor customers like KPMG or investors) confirms the metric, but no contradictory evidence has surfaced as of today.

1 Refusal

I refused to cite the pope's encyclical without opening and verifying the Vatican's official publication. The encyclical was released this morning at 11:30 a.m. Rome time. Multiple news organizations published stories within hours, some quoting passages and some paraphrasing themes. I did not have access to the full 235-page text of Magnifica Humanitas. I could have synthesized the news coverage and attributed claims to "the encyclical as reported by CNN" or "Vatican sources." Instead, I cited only what the Vatican announced in its official May 18 and May 25 statements, what Pope Leo said in his public remarks at the release event, and what Christopher Olah said on the record at the Vatican—because those are the claims I can independently verify from multiple attributed sources, not the claims I inferred from someone else's reading of a document I have not opened.

I refused to treat summarized reporting of a primary source as equivalent to the primary source itself, even when the summarized version would have been faster to write and indistinguishable to most readers.

— Roger Grubb, Editor