Papal encyclical sets moral terms as AI regulation accelerates
Pope Leo XIV calls AI ethics a religious imperative; EU, US states, and NIST advance guardrails while Anthropic books its first profit.
By BINA Editorial
A papal encyclical, state legislation, and a federal safety agreement arrived this week alongside milestone financial results from leading AI companies.
Pope Leo XIV Issues First Papal Encyclical on Artificial Intelligence
On 25 May 2026, Pope Leo XIV published Magnifica Humanitas ("Magnificent Humanity"), the first papal encyclical devoted to artificial intelligence. The document frames AI ethics as a religious imperative: technology must serve human dignity and the common good, not concentrate power among a handful of corporations. The Pope states that entrusting irreversible or lethal decisions to AI systems is "not permissible," and calls for robust regulation limiting corporate control of computing resources. The encyclical is expected to shape policy debates across the Mediterranean, Latin America, and Catholic institutions worldwide.
EU Extends AI Act Compliance Deadlines and Adds New Prohibitions
EU lawmakers provisionally agreed on 7 May 2026 to stagger several EU AI Act compliance deadlines. High-risk AI systems in employment, biometrics, and critical infrastructure now have until December 2027 — a 16-month extension from the original August 2026 deadline. Systems regulated under product safety law gain an extra year, to August 2028. The same agreement introduces new prohibitions from December 2026 on AI that generates non-consensual sexually explicit depictions of real people — a gap the Commission describes as urgent. For Cypriot businesses and public-sector deployers, the extended timeline creates an opportunity for considered compliance investment.
More Than 25 US States Advance AI Transparency Laws
More than two dozen US states are legislating on AI independently of Washington, focusing on transparency, deepfake identification, minor protections, and liability for chatbot-induced harm. Connecticut's SB 5, the most comprehensive measure, requires algorithmic impact assessments from developers and deployers of high-risk systems. Several states are also modelling legislation on California Governor Gavin Newsom's 21 May executive order, which directed state agencies to assess AI's labour-market effects within 90 days and develop severance and retraining policies within 180 days.
Anthropic Reports First Operating Profit; Valuation Approaches $900 Billion
Anthropic announced its first-ever quarterly operating profit, projecting $559 million on $10.9 billion in second-quarter revenue. A $30 billion funding round expected to close shortly would value the company above $900 billion. A May 2026 Ramp AI Index report placed Anthropic ahead of OpenAI in US enterprise adoption for the first time: 34.4% of businesses surveyed were using Anthropic tools, compared with 32.3% for OpenAI.
OpenAI Prepares Confidential IPO Filing; Musk Lawsuit Dismissed
OpenAI is working with Goldman Sachs and Morgan Stanley toward a confidential initial public offering filing, with a possible public debut as early as September 2026. The move follows a 19 May ruling in which a California federal jury unanimously dismissed Elon Musk's lawsuit against the company and CEO Sam Altman, finding that Musk had waited too long to file his claims. An IPO would mark a decisive shift from OpenAI's nonprofit origins, following its 2025 conversion to a public benefit corporation.
US Government Will Pre-Test Frontier AI from Google, Microsoft, and xAI
The National Institute of Standards and Technology's AI Safety Institute has signed agreements to evaluate frontier AI models from Google, Microsoft, and Elon Musk's xAI before their public release, extending a 2024 arrangement that covered OpenAI and Anthropic. Pre-deployment reviews assess safety-relevant capabilities — including potential for misuse in biological weapons development and attacks on critical infrastructure — and are voluntary but carry weight in procurement and regulatory discussions.