The Week The Men Who Built AI Told The World To Be Afraid

Something rare happened this past week. The very people who build artificial intelligence — the engineers, the CEOs, the Nobel laureates who fathered this technology — stood before the world's cameras and microphones and said, in different words but the same trembling tone: we may have built something we cannot fully control.

Not a pastor warning of end times. Not a conspiracy theorist on a street corner. The scientists themselves.

Let me take you through it, plainly, the way it actually happened.

The resignation that started the storm

On a Tuesday in September 2026, a young researcher named Jacob Coxon walked away from his job at Anthropic — one of the most respected AI companies on earth, the maker of the Claude chatbot. He did not leave quietly. He posted on social media that Anthropic and its rival OpenAI were, in his words, "racing straight to self-improving superintelligence and gambling with our lives." He said the people building this technology "earnestly believe that it could kill us all by the end of the decade" — and that this was not a marketing stunt to drum up attention, but a private fear many of them carry.

Hours later, Evan Hubinger, the man who leads Anthropic's alignment science team — the department whose entire job is to make sure AI does not turn against humanity — replied publicly. He did not deny it. He confirmed it.

"We really do earnestly believe AI could kill all humans," Hubinger wrote. "I personally think it is greater than 10% within the next decade... we do not yet have a plan to solve alignment for superintelligence, and are not clearly on track to."

Read that again, my brother, my sister. A senior scientist at one of the top AI labs in the world put the odds of human extinction from AI at better than one in ten — and admitted his own company has no working plan to stop it.

He was not alone. Samuel Marks, another Anthropic lead, added that fear tends to grow, not shrink, the more senior you become inside these companies. At OpenAI, a safety researcher named Marcus Williams put his own estimate at 70% within three years if nothing changes. Paul Christiano, who now sits on OpenAI's safety board, said he sees "a meaningful risk of catastrophic and irreversible loss of control in the very near term."

The godfather adds his voice

Then Geoffrey Hinton spoke. This is not a random critic. Hinton is the British-Canadian scientist often called the "Godfather of AI" — the man whose research in the 1980s laid the foundation for the very technology now worrying everyone, including him. He won the Nobel Prize for it in 2024. He left his job at Google in 2023 for one reason only: so he could speak freely about the dangers of what he helped create.

Speaking to LBC's Andrew Marr in the UK, Hinton said artificial intelligence could wipe out all jobs within twenty years. Not some jobs. All jobs — starting with what he calls "mundane intellectual labour": call centre staff, paralegals, translators, customer service agents. He has told other interviewers he now believes a 10% chance of AI wiping out humanity within a decade is "not unreasonable." His advice to young people watching their career paths disappear? Learn a trade with your hands. Train to be a plumber.

The billionaire who says the future won't need money at all

Elon Musk, never a man to whisper when he can shout, gave The Economist a different kind of warning — one dressed as good news. In a lengthy interview with editor-in-chief Zanny Minton Beddoes, Musk said he expects AI to exceed the combined intelligence of the entire human race within about five years. Within ten years, he said, humans will no longer be in control of the world. He calls this an "age of amazing abundance," where robots and AI produce so much that money becomes almost meaningless.

But sit with the second half of that sentence: humans will no longer be in control. Musk says this the way another man might describe good weather coming. Whether you find that thrilling or terrifying may depend on how much trust you place in the hands steering the wheel when we get there.

Enter the politicians — and the dragon in the room

Naturally, Washington could not stay silent. Anthropic's own chief executive, Dario Amodei, published a lengthy essay titled "We Must Pace the Frontier," arguing that AI companies should deliberately slow down their race toward ever-smarter systems. Sam Altman of OpenAI and even Musk endorsed the call. Amodei went further, warning that a "Chinese lead in AI would pose grave danger for the United States and the world," and pushed for continued restrictions on selling advanced chips to China.

Beijing did not take this lying down. Chinese Foreign Ministry spokesman Guo Jiakun fired back at a press briefing: "Fearmongering, confrontation and vicious competition will only disrupt the process of global AI governance and serve the interests of no one." China's state-run Global Times went further, accusing Amodei of running a "silent AI Cold War" dressed up as safety concern.

Meanwhile, in the White House, President Donald Trump waved off talk of AI-driven extinction altogether. His worry is different: that if America slows down, China wins the race outright. "I have concerns that if we don't win AI, we're going to be put in a very bad position," he said, putting the current American lead at roughly one year — which, in his words, "is a lot."

So here we stand: scientists warning of extinction, a president worried about losing a race, and two superpowers accusing each other of using fear as a weapon. Nobody, it seems, agrees on what the danger actually is — only that something big and unpredictable is coming fast.

Bill Gates: the danger is not just death, it is neglect

While all this fire and thunder filled the airwaves, Bill Gates offered a quieter but no less serious warning. His foundation released its tenth annual Goalkeepers Report this week, with a blunt title: Make This Matter: AI, Equity, and the Choice We Can't Delay. Its central message is this — while the rich world argues about whether AI will kill us all, it is already leaving billions of poor people behind.

The numbers Gates cites should embarrass every government that claims to care about fairness. More than 90% of the data used to train the earliest large AI models came from English-language sources. Leading AI speech systems make mistakes less than 6% of the time in English — but get things wrong more than 60% of the time in Yoruba, a language spoken by tens of millions of West Africans. As Gates put it in the report, AI sharpens every month for the roughly one billion people, mostly English speakers, who already have access — and stays frozen in place for the other seven billion.

To close that gap, at least in a small way, the Gates Foundation pledged $1 billion over the next two years toward AI tools for health workers, farmers, and teachers in the world's poorest countries — 40% to education, 40% to healthcare, 10% to agriculture, and 10% to digital infrastructure, including building datasets in languages the big models currently cannot understand. Gates Foundation chief executive Mark Suzman called it "a down payment," admitting the figure is small next to the roughly $725 billion Big Tech plans to spend on AI infrastructure this year alone. It is a fair question whether $1 billion can shift a $725 billion industry. But it puts a number, and a two-year clock, on a promise — which is more than most philanthropy offers.

Now, let me steelman the other side

It would be dishonest of me, as a writer who values truth over drama, to leave you thinking every serious voice agrees the sky is falling. It does not.

Many respected researchers argue these extinction estimates are guesswork dressed up as science — a way for insiders to sound humble and thoughtful while their companies keep racing ahead regardless. Not every employee at these labs agrees with the doom talk either; one OpenAI life sciences researcher publicly pushed back, insisting AI "has been and will continue to be an extremely beneficial technology to humanity." Critics also point out that fear sells attention, funding, and regulatory favour — a company that says "our product might end the world" is, strangely, still selling that same product to millions of paying customers the very next day. There is a real tension between the safety message and the business model, and reasonable people are right to notice it.

Some economists also argue Hinton's job-loss warnings echo predictions made about computers, tractors, and the printing press — technologies that also destroyed old jobs while quietly creating new ones nobody had imagined yet. History does not always repeat, but it does not always end in ruin either.

What this means for us, here in Ghana and across Africa

My people, here is where I put down my pen and speak from the heart of Kumasi and Accra, not from a foreign newsroom.

While the billionaires argue about whether machines will end humanity in ten years, our more urgent question is simpler and closer to home: will AI even speak our languages in ten years? Will it understand Twi, Ga, Ewe, Dagbani, Hausa — or will we remain, as Gates's own report shows for Yoruba, an afterthought coded in as a footnote while the wealthy world races ahead unbothered?

This is not a call to reject AI out of fear. Our young people at Accra Technical University and across our universities need to master this technology, not run from it. But we must also demand a seat at the table where these decisions are made — because right now, the conversation about AI's future is happening entirely in San Francisco, Beijing, and London, while Africa listens from the outside, hoping the crumbs of a billion-dollar pledge reach us before the technology has already moved three steps ahead.

A word from scripture

In all of this noise — the billionaires, the politicians, the scientists arguing among themselves — I am reminded of a word from the book of Proverbs that speaks directly to men who build towers without first counting the cost:

"Where there is no counsel, the people fall: but in the multitude of counsellors there is safety." — Proverbs 11:14 (KJV)

That is the whole story of this week in one verse. The men building AI are, by their own admission, moving forward without full counsel, without a finished plan, without safety worked out. Wisdom has never been found in one voice racing ahead alone — not one company, not one nation, not one billionaire's timeline. Safety, scripture tells us, is found in many counsellors working together, humbly, honestly. That is the very thing missing right now between Washington and Beijing, between the safety teams and the sales targets, between Silicon Valley and the seven billion people the Goalkeepers Report says AI has forgotten.

Author's note

I wrote this piece not to frighten you, but to inform you plainly, the way I would explain it to my own brother over a bowl of fufu. The people building the most powerful technology in human history are themselves unsure where it leads. Some see extinction. Some see abundance. Some see a race to be won at any cost. What all of them agree on, whether they say it directly or not, is this: the next two to five years will decide a great deal. We cannot afford to watch from the sidelines and hope it turns out well for us.

About the Author
Chief Tutu Baffour Asare Brownsy Williams is an author, filmmaker, opinion columnist, and founder of Brownsy Silva Company, a multi-disciplinary creative platform spanning novels, films, columns, and engineering content. He writes on global affairs, technology, and society for Modern Ghana, with a readership spanning Ghana and the diaspora communities of the UK, USA, Canada, and Germany. He is currently a student of engineering at Accra Technical University.

Sources: Bloomberg, NBC News, NPR, Forbes, GeekWire, Gates Foundation Goalkeepers Report 2026, The Economist, LBC, Channel 4 News, and public statements from Anthropic and OpenAI researchers on X (formerly Twitter).