Will AI Kill Us All? Doomsday Clock Expert Compares It to Manhattan Project

Will AI Kill Us All? Doomsday Clock Expert Compares It to Manhattan Project

**The Bulletin of the Atomic Scientists Just Moved the Clock to 85 Seconds to Midnight — And They're Saying the Quiet Part Out Loud About AI**

---

## The Question Nobody Wanted to Ask

Let me tell you about a guy named Daniel Holz. He's a physicist at the University of Chicago. He runs something called the Existential Risk Laboratory. And for years, he's been part of a small group of scientists who decide where to set the hands of the Doomsday Clock — that symbolic measure of how close humanity is to destroying itself.

In January 2026, Holz and his colleagues did something they'd never done before. They moved the clock to **85 seconds to midnight** — the closest it's ever been in the nearly 80-year history of the symbol . And for the first time, artificial intelligence was named as a primary driver of that move.

But here's the thing that makes Holz different from the typical AI doomer. He's not some Silicon Valley executive trying to sell you a book or a product. He's not a venture capitalist betting on regulation. He's a scientist who spent his career studying existential threats — nuclear war, climate change, pandemics.

And when he talks about AI, he doesn't reach for science fiction. He reaches for history.

"The scientists who worked on the Manhattan Project, which built the world's first atomic bombs, used the clock to warn the public and policymakers that the technology would dramatically improve and spread," Holz told the Miami Herald in September 2026 .

"It's an extremely powerful technology, just like the power of the atom was a new, extraordinary, powerful technology," he said. "And it's very important to be prudent when the scientists who are developing it are telling you 'maybe we should take a break?'" 

That's the comparison that should stop you in your tracks. Not because AI is literally a nuclear weapon. But because the people building it are saying the same things the Manhattan Project scientists said in 1945.

*This is too powerful. We don't fully understand it. And maybe we should slow down.*

---

## The 10% Number That Changed the Conversation

For years, AI safety advocates have been warning about existential risk. Most people ignored them. They were dismissed as alarmists, doomers, or worse — people with a financial interest in slowing down competitors.

Then September 8, 2026 happened.

Jacob Coxon, a 27-year-old researcher at Anthropic — one of the leading AI companies — resigned. And he didn't go quietly. He posted on X that Anthropic and OpenAI were **"racing straight to self-improving superintelligence and gambling with our lives"** .

The people building AI, he wrote, "earnestly believe that it could kill us all by the end of the decade" .

That post went **viral**. Over 170 million views. Headlines everywhere. And suddenly, the question that had been confined to niche forums and academic papers was being asked on CNN, in Congress, and at kitchen tables across America .

Then the numbers started coming out.

Evan Hubinger, Anthropic's Alignment Science lead, said there was a **greater than 10% chance** that AI could "kill all humans" within the next decade . Marcus Williams, who works on OpenAI's safety oversight team, put the risk at **70% within three years** without regulation or a slowdown .

Geoffrey Hinton, the Nobel Prize-winning computer scientist often called the "godfather of AI," told the BBC that a **10% chance of AI wiping out humanity within a decade was "not unreasonable"** .

And Daniel Holz? His reaction wasn't horror. It was relief.

"You would think my reaction would be, 'Oh no that's terrible,'" he said. "Instead, my reaction was like 'This is terrific. Finally, we can start having the conversation globally about these risks'" .

---

## Why the Manhattan Project Comparison Matters

To understand why Holz reaches for the Manhattan Project, you have to understand something about the people who built the atomic bomb.

They weren't villains. They were scientists — brilliant, patriotic, terrified of what the Nazis might build first. They worked in secret. They did something unprecedented. And then, when it was done, many of them were horrified by what they'd created.

J. Robert Oppenheimer, the scientific director of the Manhattan Project, famously quoted the Bhagavad Gita after the Trinity test: *"Now I am become Death, the destroyer of worlds."*

Albert Einstein, whose letter to FDR helped launch the project, spent the rest of his life advocating for nuclear disarmament.

And in 1945, a group of Manhattan Project scientists founded the **Bulletin of the Atomic Scientists** — the same organization that now sets the Doomsday Clock. Their goal was simple: to warn humanity about the dangers of the technology they had helped create .

The parallel to AI is striking. The people building the most powerful AI systems are the ones warning about them. They're not outsiders looking in. They're the Oppenheimers and Einsteins of our generation, saying: *We built this. We know what it can do. And we're scared.*

"It's an extremely powerful technology, just like the power of the atom was a new, extraordinary, powerful technology," Holz said. "And it's very important to be prudent when the scientists who are developing it are telling you 'maybe we should take a break?'" 

---

## What the Doomsday Clock Actually Says About AI

The Doomsday Clock isn't a prediction. It's a warning. It's a way of communicating to the public and policymakers that the risks we face are real, urgent, and interconnected.

In its 2026 statement, the Bulletin's Science and Security Board wrote that **"artificial intelligence continues to be a significant and disruptive technology"** and that its risks are growing .

The specific concerns they listed are worth reading carefully, because they're not about robots turning on their creators. They're about something more subtle — and arguably more dangerous.

**AI in military systems.** The Bulletin noted that AI is increasingly applied to **command and control, operational planning, logistics, autonomous systems, and cybersecurity** . While some of these applications are relatively benign, others are deeply concerning. The head of U.S. Strategic Command said that while a human will always make the final decision on nuclear weapons, **it's conceivable that AI will be embedded in decision-support systems used for nuclear weapons** .

Think about that. Even if a human is "in the loop," what does that mean if the human is relying on a black-box system that makes recommendations they don't fully understand? What happens when the AI says "launch" and the human has seconds to decide?

The Bulletin called this **"a serious danger"** .

**AI-driven disinformation.** This is the threat that gets less attention but may be the most insidious. The Bulletin warned that **AI is "supercharging mass disinformation"** and making it harder to address other existential threats .

Maria Ressa, a Nobel Peace Prize laureate who participated in the announcement, put it bluntly: **"We are living through an information Armageddon that's brought about by the technology that rules our lives, from social media to generative AI. None of that tech is anchored in facts. Your chatbot is nothing but a probabilistic machine"** .

The concern isn't just that AI can generate fake news. It's that AI can generate **convincing** fake news at **near-zero cost**, flooding the information ecosystem with content that makes it impossible for people to know what's true. And when people can't agree on basic facts, they can't solve problems — including the problems that could end civilization.

**AI-designed pathogens.** This is the scenario that keeps biosecurity experts up at night. The Bulletin noted that **"some researchers are concerned that AI will be employed in the design of unique new pathogens"** .

Steve Fetter, a member of the Bulletin's Science and Security Board, elaborated: **"One clear risk is the use of AI to design novel pathogens—particularly pathogens that don't exist in nature and for which no countermeasures exist"** .

This isn't science fiction. AI systems are already being used to predict protein structures — a capability that could, in theory, be turned toward designing biological agents. The same tools that could cure diseases could be used to create new ones.

---

## The Skeptics: "This Is Not Science"

Not everyone is buying the doomsday narrative. And the skeptics aren't just random people on the internet. Some of them are among the most powerful people in tech.

**Jensen Huang, CEO of Nvidia** — the company that makes the chips powering most AI systems — has been openly dismissive of the extinction warnings.

"2030 is not going to be the end of the world," he said on a podcast in September 2026. "There is 0% chance that's going to be the end of the world. Scaring people is unnecessary. It is irresponsible" .

Huang went further, suggesting that the doomsday narrative is a **strategic play** by AI companies to avoid regulation.

"Don't let this doomsday narrative cause somebody to relieve them of the laws that currently exist," he said. "Read between the lines. They're actually not asking for more laws. They're asking to be relieved of the laws we do have. And I think that that's a problem" .

That's a serious accusation. Huang is essentially saying that the people warning about AI's dangers are doing it to **avoid accountability** — to convince lawmakers that the technology is so powerful and so dangerous that only its creators can be trusted to manage it.

**President Donald Trump** has also dismissed the fears. When asked by reporters whether he had concerns about AI causing human extinction, his answer was blunt: **"No, I don't have any"** .

His concern, he said, was losing the AI race to China. "I have concerns that if we don't win AI, we're going to be put in a very bad position," he said .

That's the geopolitical argument in a nutshell: **We can't slow down because China won't.** If the U.S. pauses AI development, China will surge ahead. And whoever leads in AI will have an insurmountable advantage.

**Even some AI researchers** are skeptical. Chris Hayduk, a life sciences researcher at OpenAI, wrote that AI **"has been and will continue to be an extremely beneficial technology to humanity"** and that the debate should be about "how many billions of lives it will save" .

And The Atlantic, in a nuanced piece published in September 2026, argued that **"maybe the truth of this AI moment sits somewhere in the middle"** — that both the doomers and the skeptics are partially right, and that the format of the debate (happening on platforms that reward sensationalism) is preventing any hope of shared understanding .

---

## Frequently Asked Questions

**Q: What is the Doomsday Clock?**

A: The Doomsday Clock is a symbolic measure created in 1947 by the Bulletin of the Atomic Scientists — an organization founded by Manhattan Project scientists. It represents how close humanity is to self-annihilation. Midnight represents catastrophe. The clock is currently set at **85 seconds to midnight**, the closest it's ever been .

**Q: Why did the Doomsday Clock move forward in 2026?**

A: The Bulletin cited multiple factors, including nuclear weapons threats, climate change, and **disruptive technologies — particularly artificial intelligence**. The specific AI concerns include military applications, disinformation, and the potential for AI-designed pathogens .

**Q: Who is Daniel Holz?**

A: Daniel Holz is a physicist at the University of Chicago, founding director of the University's Existential Risk Laboratory, and chair of the Science and Security Board of the Bulletin of the Atomic Scientists — the group that sets the Doomsday Clock. He has compared the current AI moment to the Manhattan Project .

**Q: What did Jacob Coxon say?**

A: Jacob Coxon, a former researcher at Anthropic, resigned in September 2026 and posted on X that AI companies were **"racing straight to self-improving superintelligence and gambling with our lives."** He said the people building AI "earnestly believe that it could kill us all by the end of the decade" .

**Q: What is the actual risk of AI killing everyone?**

A: Estimates vary widely. Anthropic's Evan Hubinger said there's a **greater than 10% chance** within a decade . OpenAI's Marcus Williams said **70% within three years** without regulation . Nvidia's Jensen Huang said **0%** . There is no scientific consensus.

**Q: How could AI actually kill us all?**

A: The most commonly cited scenario is through **AI-designed bioweapons** — novel pathogens that don't exist in nature and for which no countermeasures exist . Other scenarios include AI in nuclear command and control systems, and AI-driven disinformation that prevents society from addressing other existential threats .

**Q: What is the Manhattan Project comparison about?**

A: The Manhattan Project was the U.S. program that developed the first atomic bombs during World War II. Many scientists who worked on it later warned about the dangers of nuclear technology. The comparison suggests that **the people building the most powerful AI systems are the ones warning about them** — just like the atomic scientists did .

**Q: What does Jensen Huang say about AI extinction risks?**

A: Huang, CEO of Nvidia, has dismissed the warnings as **"irresponsible"** and accused AI companies of using the doomsday narrative to **avoid regulation**. He said there's a **"0% chance"** AI will end the world by 2030 .

**Q: What is Congress doing about AI safety?**

A: Two bills have been proposed. The **FRONTIER Act** would set a framework for deploying advanced AI models . The **Ban Artificial Superintelligence Act** would pause advanced AI development until safety rules are established . Neither has passed.

**Q: What does the European Union say?**

A: The EU has called for **global rules** on AI safety. The European Commission's tech chief said EU law already requires companies to assess the risk of losing control of AI models, but **"that is not the case globally"** .

**Q: Should I be worried about AI?**

A: This article is not offering personal advice. The honest answer is: **experts disagree**. Some of the people closest to the technology are genuinely terrified. Others — including some of the most powerful people in tech — say the fears are overblown. What's clear is that AI is a powerful, rapidly advancing technology with real risks that deserve serious attention.

---

## Conclusion: The Conversation We Can't Avoid

Here's what I keep coming back to when I think about Daniel Holz and the Doomsday Clock.

The clock was created by scientists who had just built the most destructive weapon in human history. They weren't trying to scare people. They were trying to **warn** them. To say: *This is what we've made. This is what it can do. We need to be careful.*

Eighty years later, the same warning is being issued about AI. And the people issuing it aren't outsiders. They're the ones building the technology.

Jacob Coxon quit his job to sound the alarm. Evan Hubinger, who still works at Anthropic, says there's a 10% chance his own company's products could kill everyone. Geoffrey Hinton, who won a Nobel Prize for his work on neural networks, says the risk is real.

These aren't people with nothing to lose. They're people who have everything to lose — and who are speaking up anyway.

But they're also not the only voices in the room. Jensen Huang says the warnings are a regulatory power play. Donald Trump says the only thing that matters is beating China. And millions of people are using AI every day to write emails, generate images, and automate work — with no apparent catastrophe.

The truth is probably somewhere in the messy middle. AI might not kill us all. But it's clearly a technology with enormous power — power that could be used for great good or great harm. And the people building it are telling us they're not sure which way it will go.

That's the conversation Holz wants us to have. Not a debate about whether AI is good or evil. But a serious, global conversation about **how to manage a technology that could reshape civilization**.

The Doomsday Clock isn't a prediction. It's a warning. And right now, it's saying: **Pay attention. This is real. And we need to talk about it.**

---

## Disclaimer

**This article is for informational and educational purposes only. It does not constitute investment, policy, or personal safety advice. The author has no positions in any AI-related securities mentioned. Information presented here is based on publicly available sources and reported statements as of the publication date. The debate about AI existential risk is ongoing, and expert opinions vary widely. Readers should seek diverse perspectives and consult qualified professionals before making decisions based on the information presented. The author does not have access to non-public information about AI systems or their capabilities.**