Vatican, Anthropic Urge AI Restraint.

Mazen Al Sabbagh

— 6 min read
Share

TL;DR

  • Papal Mandate: Pope Leo XIV's "Magnifica Humanitas," a 42,300-word encyclical, implores governments and corporations to decelerate AI development, presented alongside Anthropic co-founder Christopher Olah who acknowledged internal industry incentive conflicts.
  • China's Talent Clampdown: Beijing restricts overseas travel for key AI researchers at firms like DeepSeek and Alibaba, demanding prior approval, a clear escalation to safeguard its technological lead.
  • Altman's Revision: OpenAI CEO Sam Altman walked back prior warnings about AI-driven job displacement, expressing "delight" that the expected "jobs apocalypse" for white-collar entry-level roles had not materialized as rapidly.
  • Huawei's Chip Gambit: Huawei unveiled "LogicFolding" chip architecture, claiming it can achieve 1.4-nanometer-equivalent density within five years without advanced lithography, positioning itself to challenge current industry roadmaps.
  • Anthropic's Valuation Surge: Anthropic's latest funding round, exceeding $30 billion, is poised to close this week, valuing the company above $900 billion and positioning it as the most valuable AI startup.

Lead Story: Vatican and Anthropic Advocate for AI Temperance

Pope Leo XIV issued his inaugural encyclical, "Magnifica Humanitas," delivering the most comprehensive institutional critique of the AI sector to date. Released Sunday from the Vatican, the 42,300-word document frames artificial intelligence as a civilizational inflection point akin to the Industrial Revolution, a parallel underscored by its May 15 signing date, precisely 135 years after Leo XIII's "Rerum Novarum" on workers' rights. The encyclical urges governments to moderate AI development velocity, advocates against exclusive private control of AI data, and calls for the technology's "disarmament" from military, economic, and cognitive competition. Leo's text explicitly warns that certain autonomous weapons systems now operate "beyond human control," marking a clear papal renunciation of modern just war theory.

The event's distinctiveness was magnified by Christopher Olah, Anthropic's co-founder and head of interpretability research, who presented alongside the Pontiff. Olah stated: "Every frontier AI lab, including Anthropic, operates inside a set of incentives and constraints that can sometimes conflict with doing the right thing," advocating for religious communities, academics, and governments to act as "informed critics."

This backdrop provides a sharper context. The Washington Post reported Olah's Vatican appearance coincides with Anthropic's ongoing lawsuit against the Trump administration, which rescinded a $200 million Pentagon contract and designated the company a "supply chain risk" after its refusal to permit mass surveillance or fully autonomous weapons applications. The Daily Caller highlighted the co-founder central to this Pentagon dispute was now influencing Vatican AI policy. This strategic alignment — Anthropic positioning itself with the Vatican against the White House — is deliberate. It recasts the company's safety commitments as a differentiating asset rather than a liability, precisely as it finalizes the largest private funding round in AI history.

Both The New York Times and CNN covered the release as a definitive statement of Leo's papacy. Its practical impact hinges on whether it furnishes policymakers, particularly in the EU and the Global South, with a moral framework translatable into concrete regulation.

In Other News

China fortifies AI talent pool amidst intensified tech rivalry. Bloomberg reported Monday that Chinese government bodies now require key AI researchers from private enterprises, including DeepSeek and Alibaba, to secure approval before international travel. These measures target individuals deemed "strategically important" for national AI development. The Next Web detailed instances where senior researchers are asked to surrender passports, effectively enacting an exit ban without judicial oversight, citing the potential for state or commercial secrets disclosure. This development follows Huawei's Sunday announcement of "LogicFolding" chip architecture and a "Tau Scaling Law," promising 1.4-nanometer-equivalent transistor density by 2031, circumventing EUV lithography machines currently inaccessible to China. Counterpoint Research analysts cautioned that a stacked design does not inherently resolve "full process, yield, power, thermal, and device-performance problems." However, the convergence of travel restrictions, chip innovation, and DeepSeek's reported V4 model rework for Huawei Ascend chips signals a deliberate decoupling of the Chinese AI ecosystem from American supply chains.

Altman softens stance on AI-driven job displacement. Addressing a Commonwealth Bank of Australia conference virtually from Sydney, OpenAI CEO Sam Altman stated Monday he was "delighted to be wrong" regarding the speed of AI-induced job losses. "I thought there would have been more impact on entry-level white-collar jobs being eliminated by now than has actually happened," he conveyed to CBA CEO Matt Comyn. Altman described using AI for Slack and email responses before reverting to manual handling, citing a preference for human interaction. This concession follows HSBC's CEO recently advising 211,000 staff not to "fight" AI, and Standard Chartered announcing 7,800 AI-linked job reductions. Concurrently with Altman's remarks, Cloudflare CEO Matthew Prince published a Wall Street Journal op-ed justifying a 20% workforce reduction (approximately 1,100 roles) due to AI's capacity to replace "measurers"—middle management, finance, legal, and compliance functions. The divergence between Altman's reassurances and observed corporate actions implementing AI continues to expand.

Anthropic's record funding round nears completion. Bloomberg reported Friday that Anthropic's latest funding round, exceeding $30 billion and valuing the company beyond $900 billion, is expected to finalize this week. This would establish it as the world's most valuable AI startup, surpassing OpenAI. The company has informed investors its annualized revenue run rate is projected to exceed $50 billion by the close of June. This marks Anthropic's second $30 billion-plus raise within 14 weeks, following its February Series G which closed at $380 billion.

X / Social Pulse

The Pope's encyclical dominated weekend and Monday discourse, with the Washington Post's framing – "Anthropic aligns with Vatican over White House" – solidifying as the prevailing narrative. Olah's statements garnered significant attention beyond the AI safety community, resonating within defense and policy circles given the active Pentagon contract dispute. China's new travel restrictions intensified Monday's online discussions, particularly following reports of researchers surrendering passports, which commentators noted constitutes an exit ban rather than mere approval requirements. Altman's "jobs apocalypse" retraction drew sharp criticism from recently laid-off workers, who cited Layoffs.fyi's count of over 114,000 tech job cuts in 2026.

One to Watch

Samsung's union vote concludes tomorrow morning with passage highly anticipated. Combined voter turnout reached 91.89% by Monday afternoon, with the Samsung Electronics Labor Union recording 92.74% and the National Samsung Electronics Union 85.98%. A Suwon District Court ruling Monday rejected an injunction filed by five Device Experience employees seeking to halt the vote, finding no "serious flaws" in the bargaining proposal. The Donghaeng Labor Union, representing non-chip workers, has withdrawn from joint bargaining, alleging its interests were marginalized. The compensation disparity remains stark: semiconductor employees are positioned to receive 210-600 million won in performance pay, while DX employees are allocated approximately 6 million won in company shares. Results are scheduled for May 27 at 10 am KST.

Quick Hits

  • Lenovo achieved an all-time share price high following a reported 38% profit growth and a $21 billion AI server pipeline, with shares surging 20% on five times normal volume.
  • DeepSeek deferred its V4 model release to rework its software stack for Huawei Ascend chips, per a CCTV-affiliated account, prioritizing domestic hardware integration over time-to-market.
  • Germany launched a EUR 125 million "Next Frontier AI" initiative via its federal innovation agency SPRIND, targeting the creation of Europe's own frontier AI companies.
  • Micron surged 18% towards a $1 trillion market cap after UBS tripled its price target to $1,625, citing AI-driven memory demand and long-term supply agreements mitigating historical earnings volatility.
  • Huawei has already mass-produced 381 chips utilizing its Tau Scaling techniques over six years, with the first LogicFolding-based Kirin processors anticipated this fall.

Monday's developments converge on the issue of control: who wields it, who seeks it, and who endeavors to moderate its application. The Pontiff advocates for governmental oversight of AI laboratories. Beijing moves to contain its researchers domestically, now employing passport confiscation. Altman attempts to assuage public fears regarding the job implications of his technology. Huawei seeks to demonstrate self-sufficiency from American hardware. Meanwhile, Anthropic, whose co-founder asserted that his own industry's incentives are not always trustworthy – all while his company litigates against the White House over a Pentagon contract – is finalizing the largest private funding round in AI history. The question of AI governance remains unresolved, yet the array of institutions claiming influence expanded significantly this weekend, and the inherent contradictions between their objectives are increasingly evident.

Sources

Lock in. M. mazen@thorterminal.com

Read more