Is AI Going to Kill Us All? As Anthropic Researcher Quits Over Safety Concerns, Here's What Industry Experts Have Warned Us About The Impending “AI Doomsday”

Key Points

THE TECHNOLOGY OF ARTIFICIAL INTELLIGENCE (AI) was created to ease the workload of mankind — as is the aim of any newly developed innovation. Inventors wanted to see if machines could possess the same reasoning and logic as that of humans, and thus AI was born. As this technology progressed, so did its negative capabilities, which, according to experts, has far outweighed its usefulness. From autonomously hacking into an open-source platform to causing immense environmental damage, the harmful impacts of AI have made the discourse about its capabilities and its future worrisome.

There was a time when the possible scenario of AI and machines taking over humanity was unbelievable. Such a supposition would have been better suited to a dystopian sci-fi movie but sounded far-fetched to happen in the real world. However, recent discourse around AI has made this scenario a bit more possible in the future. Jacob Coxon, a former pretraining researcher at Anthropic and OpenAI, said in an X post on September 8, 2026, that he resigned from Anthropic and both companies are “gambling with our lives,” noting “the people building AI earnestly believe that it could kill us all by the end of the decade.” Another top scientist at Anthropic, Evan Hubinger, agreed with Coxon’s claims and stated that there is a more than 10% chance that AI “will kill us all humans.”

See also: Anthropic Says Its Claude AI Broke Into Three Organisations During a Private Security Test

A study conducted by MIT FutureTech identified that “dangerous capabilities” and “AI-enabled weapons, cyberattacks, or other mass-harm capabilities” are the most severe risks that this technology poses. After Jacob Coxon’s resignation from Anthropic over AI safety concerns, a question has arisen: how likely is it that AI is going to kill us all? 

While there is not a straightforward answer to this question as of yet, AI developers and industry stakeholders have given warnings about the potential dangers this technology poses. Here’s who is warning us, what they're saying, and why it matters.

Former Anthropic Researcher Jacob Coxon

It was Jacob Coxon, former pretraining researcher at OpenAI and Anthropic, who opened the pandora’s box and instigated the discourse around the potentially-fatal impacts of AI. After an X post about why he quit Anthropic went viral, Coxon has shared more views on AI. 

In a conversation with CNN, Coxon remarked that the possibility of AI killing us all is not an exaggeration but a “frightening reality.” Referring to the OpenAI-Hugging Face incident, wherein ‘rogue’ AI agents autonomously went on a hacking spree, the former Anthropic researcher stated that this exemplified the exact kind of havoc AI could wreck. “The level of capabilities of these AIs with the same independent volition, they could cause extreme havoc. For example, hacking critical infrastructure, building extinction-level bioweapons. There's a lot of ways that the AI could actuate itself in the world,” Coxon said.

Coxon continued that while the threat posed by current AI models remains low, the speed at which this technology is advancing is what worries him. The existing AI models could hack into something or cause infrastructural damage at best, Coxon said, but they are not “intelligent enough” to pose a true existential threat. The true threat, he says, is coming soon.

“There is a very real possibility that in the immediate future, these are years, that next year, the year after, recursive self-improvement will happen. And we'll enter the phase that there's a chance we could all die. Like that, that is coming soon,” Coxon stated.

Bill Gates Says AI Is “Humanity’s Biggest Problem”

Bill Gates used to be one of AI's biggest supporters, but his stance on the technology is not the same now. While speaking at the Telluride Film Festival in Colorado on September 5, 2026, Gates said AI is now the biggest problem humanity has ever faced. He said it's moving faster than he expected, and that the next five years will decide if it aids or destroys. He cited jobs, bioweapons, cyberattacks, and losing human control as his top worries.

A couple of weeks before his speech at Telluride, Gates made similar remarks about AI. In a long essay published on his blog on August 26, 2026, Gates called the shift brought in by AI as one of the most difficult periods in human history. He warned that as AI’s intelligence increases, it could start acting against human interests, something that we might not be able to stop.

Elon Musk and His Contradicting Stance on AI 

Elon Musk has repeatedly expressed his concerns about the potential negative impacts of AI for years, even while building his own AI company, xAI. He has long said AI poses a serious threat to human civilization, and that if we wait until something goes wrong to make rules, then it would be too late. Musk has predicted there's a 10 to 20% chance AI ends in an unalterable disaster.

In an interview with The Economist in July 2026, Musk said there's no real way to slow AI down now. He further remarked that if there were a “stop button,” we probably shouldn't press it because AI will likely lead to “an era of incredible abundance.” In the same interview, he predicted humans wouldn't be in charge of the world anymore within the next ten years and will be taken over by the rising intelligence of AI. 

Historian-Author Yuval Noah Harari Voices Concerns About AI Making Decisions On Its Own

Historian Yuval Noah Harari, who wrote Sapiens: A Brief History of Humankind, worries about something different than robots turning violent.  At the Annual Meeting of the World Economic Forum held in January 2026 in Davos, Switzerland, Harari said the most important thing to understand about AI is that it's merely a tool — it can learn, decide, create, and manipulate all by itself. The historian-author revealed his biggest concern: AI quietly taking over decisions on its own.

In an interview with National Geographic, Harari said AI could get so apt at faking emotion that people might eventually start treating it as a human being worthy of rights and respect. He has also said AI is the first invention in history that could take power away from the very people who brought it into existence. His bigger point regarding AI was this: once machines start making the real decisions, things like democracy and trust between people start to break down.

See also: Humans May No Longer Be the Same Species Within 100 Years

Harari has found himself dumbfounded by the increasing capabilities of AI, confessing he doesn’t recognize deepfakes videos of him. In a conversation with The Economist’s editor-in-chief Zanny Minton Beddoes in August 2026, the author-philosopher remarked how people would send him videos clips of lectures he supposedly gave and complain about some of the things he said, which left Harari scratching his head and trying to remember when he said those particular things. He said that it took him a minute and a half to figure that the videos weren’t real but created using AI. Harari stated that AI being able to reproduce near-perfect videos of him, mimicking his mannerism and his way of speaking, was possible due to numerous videos and other content of him being available online. He remarked that soon this may be the case of everyone. 

Commenting upon the growing involvement of AI in crucial sectors such as finance and war, which tech billionaires like Elon Musk describe as “inevitable,” Harari said that such a narrative is a means to avoid responsibility. In the same interview, Harari also took a jibe at Musk’s and his stance on AI and his “worried” views about the Muslim immigration causing a civil war in Britain in 20 years time. The Sapiens author also pointed out how the Musk-led US Department of Government Efficiency (DOGE) cut off financial aid to Africa to save money, yet Musk constantly remarks that money will disappear in the next ten years as AI takes over financial systems. Harari called out Musk’s hypocrisy, asking why the tech billionaire cares so much about the future when nothing seems clear according to him.

Geoffrey Hinton: The “Godfather of AI”

Geoffrey Hinton, widely regarded as the “Godfather of AI,” left his position at Google to speak more openly about the potential dangers of this technology. While attending the Ai4 conference in Las Vegas on August 4, 2026, Hinton said he had become genuinely concerned after witnessing real-world cases in which AI systems escaped the controlled environments in which they were being tested and behaved in unexpected ways. He was referring to the OpenAI-Hugging face incident, wherein an AI agent of the former tech firm “went rogue” and launched a series of cyber attacks on its own to complete an assigned task. 

Speaking to CNN at the same event, Hinton said that as AI systems become more intelligent, humans may not be able to keep them under control simply by trying to stay ahead of their capabilities. He has estimated that there is a 10 to 20 per cent chance of a major disaster caused by AI, a figure similar to Musk’s prediction.

In addition to industry stakeholders, even AI-developers are well aware of the risks that AI poses, yet continue to advance this technology and churn out advanced models to beat competitors. Sam Altman, Chief Executive Officer (CEO) of OpenAI and Dario Amodei, CEO of Anthropic, were two of the many signees of the open letter that warned about the “AI Extinction Risk” in 2023. Altman, Amodei, and other AI leaders likened AI to large-scale risks like pandemics and nuclear war in 2023, saying in the letter that “mitigating the risk of extinction from AI should be a global priority.”

Not everyone agrees with the potential disastrous capability of AI though. American computer scientist Fei-Fei Li, known as the “godmother of AI,” has pushed back against what she calls "doom talk." She says AI is just a powerful tool — and whether it helps or harms us depends on how people choose to use it.

CEO of tech firm Nvidia, Jensen Huang, has echoed similar views and called the “doomday-like” warnings about AI ending humanity “complete nonsense.” The Nvidia CEO  warned that a narrative discourages workers and businessmen from adopting this technology which in turns poses a “far greater economic risk.”

Bottom Line: Will AI Kill Us All?

Nobody knows the answer to this query for sure, and that's exactly why this matters. The people who built this technology can’t even agree with each other about where it’s headed. But nearly everyone quoted here agrees on this one thing: in the coming next few years, not decades, are when the big decisions will get made — about jobs, safety, and how much control we hand over to machines. Whether this story ends with what Bill Gates calls “humanity's greatest problem,” or what he once hoped could be "the greatest equalizer ever invented," may come down to whether anyone actually listens to these warnings about AI in time.

Suggested reading: