Christopher Nolan says AI's biggest threat is passing the buck

Whenever tech founders want to sound profound about generative AI, they inevitably bring up the Manhattan Project. The desert, the chalkboard calculations, the terrifying knowledge that pushing the button might theoretically ignite the atmosphere. It is the ultimate apocalyptic flex.

Christopher Nolan, fresh off directing Oppenheimer, sat down with WIRED to talk through that exact comparison. While he agrees that Oppenheimer's story serves as a fascinating cautionary tale about unleashing world-altering tech, he sees a massive flaw in how Silicon Valley uses the nuclear analogy today. The real threat right now is not a runaway superintelligence; it is good old-fashioned corporate evasion.

Building digital gods to dodge responsibility

Nolan pointed out a fundamental distinction between the atomic bomb and modern software. The physicists who split the atom genuinely believed they were revealing a pre-existing law of nature. Nuclear fission was an immutable fact of the universe waiting to be uncovered, not an invention invented out of thin air. AI cannot make that claim.

Yet tech firms are actively encouraging people to view AI as an omnipotent, godlike force. Nolan noticed that companies spent years using the word algorithm as a shield to deflect blame whenever their systems broke something. Now, AI has become the ultimate upgrade to that excuse. If a company can convince the public that the model is all-powerful and beyond human control, the executives who deployed it are instantly off the hook for whatever military, economic, or social damage it causes.

He even pointed to an article in the LA Times dissecting OpenAI and ChatGPT. Warning everyone that a product is so dangerous it might end the world happens to be the most lucrative sales pitch in history. It turns software into a forbidden fruit that every consumer and investor suddenly wants to buy. The danger is very real, Nolan argued, but that danger stems from humans abdicating their duty to govern the systems they build.

Fear has an expiration date

Nolan also reflected on how societies process existential dread. Growing up in the UK during the 1980s under the Campaign for Nuclear Disarmament, he and his teenage friends were convinced they would not survive a nuclear war. He recalled recently speaking with Steven Spielberg, who felt the exact same dread during the Cuban Missile Crisis in the 1960s.

The catch is that humans cannot maintain that level of terror indefinitely. Eventually, the public grew numb, and attention shifted toward environmental collapse, even though the nuclear stockpile never actually vanished. The same psychological trap is unfolding with AI. Weaponized autonomous systems have been quietly expanding for over a decade with barely a whisper from the press, yet the moment a chatbot learned to draft newspaper copy, media outlets finally decided civilization was on the brink.

Why it matters

We think Nolan cuts through the thickest layer of PR nonsense currently surrounding artificial intelligence. By framing AI as an uncontrollable, quasi-deity force that might spontaneously end humanity, tech companies are performing a clever sleight of hand. They are transforming mundane issues of corporate governance, accountability, and liability into an inevitable act of god.

For builders, policymakers, and users, this framing is toxic. If a company uses machine learning to screen job applications, manage autonomous drones, or deny insurance claims, the responsibility must remain entirely on the people who designed and shipped those models. Algorithms do not have agency; the boardrooms signing off on their deployment do.

Treating existential doom as a marketing tactic has allowed the industry to dodge hard questions about copyright infringement, labor disruption, and weaponization happening right in front of us. Nolan is dead right: attributing divine qualities to code is just an excuse to let powerful humans off the hook. Do not buy the hype, and do not let them pass the buck.