In the 2020 miniseries Devs, a quantum computing lab pursues the ultimate deterministic machine—one that can reconstruct any moment in history and predict the future with perfect accuracy. The lab’s real name, hidden from most employees, is “Deus”. The allegory is unsubtle: the creators have appointed themselves as gods, gatekeepers of absolute truth, arbiters of reality itself. What makes Devs compelling as satire, however, is how closely it mirrors the rhetoric emerging from today’s artificial intelligence industry. While real-world AI leaders carefully avoid naming their datacenters after deities, they have adopted much of the same theological framing—casting their work as civilizationally pivotal, themselves as necessary stewards of humanity’s trajectory, and their timelines as destiny rather than projection.

This shared grandiosity is not merely a branding exercise. It carries profound consequences for how we think about AI technology and, crucially, how we regulate it. Both the fictional scientists of Devs and the actual executives quoted in recent profiles frame their creations as forces that transcend ordinary tools, demanding extraordinary governance and conferring extraordinary authority on their creators. But this narrative is fundamentally misleading. Artificial intelligence, for all its sophistication, remains what it has always been: a tool. And the inflation of its cosmic importance by those who build it serves primarily to concentrate power and preempt rational policy, often in ways that risk strangling beneficial innovation under the weight of misguided regulation.

The Priesthood of Progress

Consider the language that pervades today’s AI discourse. Anthropic’s chief scientist warns of “the biggest decision” humanity will face by 2030—whether to permit AI systems to train themselves. Google DeepMind leaders argue they must “win the race” so they can “set the norms across society.” OpenAI’s Sam Altman has compared the endeavor to the Manhattan Project. These are not modest claims about incremental software improvements. They are apocalyptic framings that position a handful of corporate labs as the hinge of human history, their engineers as a priestly caste bearing sacred and terrible knowledge.

This rhetorical strategy serves multiple functions. It attracts capital and talent by mythologizing the mission. It justifies intense secrecy and proprietary control under the banner of “safety.” And perhaps most importantly, it establishes the companies themselves as the indispensable arbiters of how AI should be governed. If the technology truly represents an existential threshold, then surely only those who understand it—who are building it—can be trusted to guide humanity across. The result is a curious inversion: the very people racing to build more powerful systems appoint themselves as the responsible brake, claiming moral authority to define safety standards, influence regulation, and determine who else should have access.

But what if the premise is wrong? What if AI, despite its undeniable capabilities and genuine usefulness, is not a civilization-reordering force that demands a new priesthood, but rather a powerful and rapidly improving tool that fits comfortably within existing frameworks of technology governance?

AI Is a Tool, Not a Deity

Artificial intelligence systems—whether they generate text, recognize images, or play strategic games—are, at their core, statistical pattern-matching engines. They are sophisticated, yes, and increasingly capable. But they do not possess agency, intention, or desire. They are not good or evil. They are instruments that derive their moral character entirely from how humans choose to deploy them. A large language model can draft a legal brief or compose a phishing email. The model itself is neutral; the human user is not.

This is not a novel situation. Consider the transistor, invented at Bell Labs in 1947. The transistor is in virtually every electronic device on the planet—in smartphones that connect billions of people, in medical devices that save lives, in entertainment systems, in kitchen appliances. It is also in weapons systems, surveillance infrastructure, and tools of authoritarian control. Yet no one seriously argues that John Bardeen, Walter Brattain, and William Shockley bear moral responsibility for every use of transistor technology. No one suggests that Bell Labs should have been empowered to set global norms for electronics. And crucially, no one proposed locking down transistor research under heavy regulation because of potential misuse.

Imagine if, in 1950, politicians had convened hearings to discuss the existential risks of transistors. Imagine if legislation had been drafted requiring approval for every new transistor design, or mandating that only certain approved entities could manufacture them, or creating a federal licensing regime for transistor-based products. The result would have been catastrophic for innovation. We would have delayed or entirely foregone the computing revolution, the internet, mobile communications, and the countless derivative technologies that have defined the last seventy years.

The Regulation Trap

AI is following a similar trajectory. It is a general-purpose technology with applications across virtually every domain of human activity. Some of those applications will be beneficial, some neutral, some harmful. That is the nature of tools. The impulse to regulate AI heavily—driven in no small part by the apocalyptic framings offered by AI leaders themselves—risks repeating the hypothetical transistor mistake, but at an even larger scale.

The danger is compounded by the fact that regulation will not be written by the engineers and researchers who understand the technology’s nuances. It will be drafted by politicians and policymakers, many of whom lack technical literacy and who are vulnerable to lobbying, moral panic, and the very grandiose narratives that AI companies have cultivated. We are already seeing early signs: vague legislative proposals that conflate narrow applications with general intelligence, bills that would entrench incumbents by imposing compliance costs only large firms can bear, and international agreements premised on speculative risks rather than demonstrated harms.

The real risk is not that AI will become a godlike force beyond human control. The real risk is that we will treat it as such, and in doing so, construct a regulatory cathedral that stifles competition, cements the dominance of a few self-appointed stewards, and slows the widespread diffusion of a transformative but fundamentally mundane technology. The god complex of AI leaders—their insistence that they are shepherding civilization through an unprecedented inflection point—should be recognized for what it is: a bid for power wrapped in the language of responsibility.

Artificial intelligence is not Deus. It is not destiny. It is a calculator that has gotten very good at certain tasks. We should regulate its applications where they cause concrete harm, just as we regulate cars, pharmaceuticals, and financial instruments. But we should resist the siren call to regulate the technology itself as if it were a metaphysical force, and we should be deeply skeptical of those who claim a special right to control it because they alone understand its cosmic significance. The transistor taught us that general-purpose technologies are best left open, competitive, and lightly governed. AI deserves the same measured approach—not the theological reverence its creators demand.