How Anthropic, OpenAI, The Vatican And Congress Want To Govern AI

Anthropic, OpenAI, the Vatican, and the U.S. Congress all agree that artificial intelligence needs guardrails. But what those guardrails should protect first—catastrophic risk, human dignity, or national competitiveness—is tearing the coalition apart.

The four institutions represent some of the most influential voices in modern AI governance. In July 2026, they each released statements or policy proposals aimed at shaping how the world's most transformative technology is regulated. The alignment ends there. Anthropic, the safety-focused lab founded by former OpenAI employees, prioritizes the risk of an AI catastrophe—a model that escapes human control. OpenAI, the creator of ChatGPT, now advocates for government licensing and safety testing while also pushing for U.S. leadership. The Vatican, through its Rome Call for AI Ethics, insists human dignity and solidarity must be the non-negotiable foundation. And on Capitol Hill, lawmakers are racing to pass bills that keep America ahead of China in the AI arms race, sometimes at the expense of strict safety measures.

This is not a new debate, but the urgency is. Generative AI has moved from labs to billions of users in less than three years. Lawmakers and religious leaders alike recognize that the window to set binding rules is closing fast. The European Union already enacted the AI Act in 2024, forcing the U.S. and other nations to catch up. Yet inside the United States, Congress remains fragmented—dozens of competing bills, no comprehensive federal AI law.

Anthropic's CEO Dario Amodei has repeatedly warned that the most advanced systems could one day pose extinction-level risks. The company's proposed governance model includes mandatory licensing for frontier models, incident reporting, and a dedicated AI safety agency. OpenAI, under CEO Sam Altman, has taken a slightly different line: it supports licensing and safety audits but also emphasizes the need to preserve American innovation and competitiveness. The Vatican, through the Pontifical Academy for Life, has gathered signatories including Microsoft, IBM, and Cisco to its Rome Call, which centers on principles like transparency, responsibility, and the protection of vulnerable populations. On the legislative front, Senate Majority Leader Chuck Schumer has driven the SAFE Innovation Act, which would allocate billions to AI safety research while boosting domestic AI capabilities.

The underlying tension boils down to a single question: what is the primary harm AI poses? For Anthropic, it's an existential tail risk. For the Vatican, it's the erosion of human dignity and an increase in inequality. For Congress, it's the threat of losing global technological primacy. These are not mutually exclusive, but they demand different regulatory priorities. Insiders say the real challenge is sequencing—what to regulate first, and how much to slow down development.

What happens next depends on whether these actors can find common ground. The Vatican has invited all parties to a high-level summit on AI ethics in Rome later this year. Meanwhile, Congress faces a midterm election and increasing pressure from the public to act on AI safety. Some observers predict a patchwork of sectoral regulations—healthcare AI, autonomous vehicles, and facial recognition—will precede a comprehensive framework. Others argue the sheer speed of AI progress will leave regulators behind. One thing is certain: the debate over AI guardrails is no longer theoretical. It is being thrashed out among the most powerful institutions on Earth, each with a different vision of what a safe AI future looks like.