When a president issues an executive order setting up voluntary federal vetting of the most powerful AI models — with the major labs deciding whether to participate — what do religious traditions and the Constitution say about who should be in charge of deciding whether a powerful new tool is safe to release?
Powerful new tools are not new in human history; what is new is the speed at which a small number of private firms can put one into the world. Religious traditions speak directly to that situation: the builder of something dangerous bears the parapet duty; the holder of a powerful entrustment is answerable for its use; trust requires accountability. The Constitution's answer to who should make the rules of general application is Congress, not the executive alone, and certainly not the regulated firms themselves. A voluntary framework is better than nothing, and the people who designed it are serious about the risks. But "voluntary" is exactly the word the religious and constitutional traditions both flag as the warning: a duty that the duty-holder may choose to skip is not, in the end, a duty.
Religious perspectives
Religious traditions do not address artificial intelligence directly, but they speak with unusual clarity about the deeper question underneath this debate: the responsibility that comes with mastering a powerful new thing, and the danger of putting the keys to it in the hands of the people most likely to benefit from saying yes.
Judaism: The Torah's building laws include the parapet — "you shall make a parapet for your roof, that you may not bring the guilt of blood upon your house, if any one fall from it" (Deuteronomy 22:8). The principle is preventive responsibility: the person who builds something dangerous bears the obligation to make it safe, before the harm occurs, not after. The same logic runs through rabbinic nezikin (damages) law: ownership of something that can hurt people brings duties, not merely opportunities.
Christianity: The parable of the talents (Matthew 25:14-30) frames every powerful capability as an entrustment — given for stewardship, not free disposal — and "to whom much is given, much will be required" (Luke 12:48). The Christian tradition's emphasis on prudence as a cardinal virtue treats the use of powerful tools not as a matter of inventor's discretion but of duty to those who will live with the consequences.
Islam: The Quranic doctrine of amana — trust — holds that human beings are entrusted with what they did not create and are answerable for its use. Classical Islamic ethics applied this directly to the introduction of new technologies and trades, requiring that those who held the new capability take responsibility for its harms — a principle that maps cleanly onto the modern question of who is accountable when an AI system causes damage.
The shared insight: Religious traditions do not oppose powerful tools — they oppose powerful tools held without responsibility. Across traditions, the test of whether a new capability is being introduced rightly is who bears the cost if it fails: the builder, who profits from its use, or the bystander, who never agreed to the risk.
Constitutional & legal framework
The Constitution does not mention artificial intelligence, but it does mention who is supposed to make rules of general application that bind private actors — and the answer is Congress, not the executive alone.
The June 2, 2026 executive order: "Promoting Advanced Artificial Intelligence Innovation and Security" directs federal agencies to develop a framework under which developers of frontier AI models may voluntarily submit their systems for federal evaluation against national-security risks — chemical, biological, radiological, nuclear, and serious cybersecurity threats. The administration framed it as a balance between addressing real misuse risks and avoiding regulation that could "stifle innovation." Critics noted what is missing: any mandatory testing, any binding pre-deployment safety standard, and any enforcement mechanism. The most powerful AI systems in the country may be released into general use after a process the labs themselves decide whether to enter.
Article I, Section 1: "All legislative Powers herein granted shall be vested in a Congress of the United States." Setting binding rules for the development and deployment of powerful new technologies — including liability, safety standards, and the consequences of failing to meet them — is a legislative function. Executive orders may direct agencies to act within authority Congress has granted; they cannot manufacture new regulatory power, and a voluntary framework is, in effect, an admission that the executive branch lacks the authority to require what it would prefer.
The federalism counterweight: States have begun acting on AI on their own — California, Colorado, and others have passed AI legislation — and the order does not preempt them, leaving a genuine patchwork. Whether a single federal standard would serve safety better is a real question, but the federalist case for state experimentation is also a real argument.
The honest other side: Mandatory regulation of cutting-edge technology can entrench incumbents (who can afford compliance), slow American competitiveness against rivals (notably China), and lock in the wrong rules early. A voluntary framework that builds shared safety practice without legislating into a moving target is a defensible policy choice — and a serious one, made by people inside the administration who care about AI risk. The constitutional concern is not that this particular policy is wrong; it is that policy of this magnitude — affecting the most powerful tools humans have built since nuclear weapons — is being shaped by executive order and lab cooperation rather than by the legislative process the Constitution names.
Sources cited:BibleConstitution