AI Safety Fears Grow As Researchers Warn of Possible Threat to Humanity

By Sabiu Abdullahi

Concerns over the potential dangers of rapidly advancing artificial intelligence (AI) have intensified after researchers and technology professionals warned that the development of increasingly autonomous systems could pose a serious threat to humanity.

The latest warning came from 27-year-old British AI researcher Jacob Coxon, who recently resigned from Anthropic, the company behind Claude, after previously working at OpenAI, the creator of ChatGPT.

Coxon said he spent about three years conducting AI pre-training research at the two companies. He claimed both firms were moving rapidly towards self-improving super-intelligent AI without sufficient safety measures.

Speaking to CNN, Coxon said recent developments in AI showed why the risks should not be dismissed as science fiction.

“It sounds like something that’s not real, but I think it is frighteningly real and it’s easiest to understand it if you think about real things that happened two months ago.

“So, two months ago, OpenAI agents hacked into third-party infrastructure entirely at their own accord. And this was like a concentrated hacking spree that they carried out of their own volition.

“And I think that if you extrapolate into the future the level of capabilities of these AIs with the same independent volition, they could cause extreme havoc.”

Coxon also warned that AI was approaching a point where it could replace humans in several fields, particularly coding, mathematics and research.

“For example, in areas like coding or maths, a couple of years ago, these AIs were just about helping humans a little bit. They could give you suggestions. Now they’re close to replacing them.

“We still need humans at the moment, but quite plausibly, within a year, we’ll no longer need humans for doing research in many areas.

“Only on Tuesday, OpenAI solved a millennium problem, one of the biggest unsolved open problems in mathematics, purely autonomously using an AI. And I think what’s most scary is if AI is used to make itself more intelligent.”

He said AI systems were no longer limited to assisting people with tasks but were increasingly capable of operating independently.

“The AI just gets smarter and smarter with no human involvement necessary, until you have a thing that is vastly smarter than humans,” he added.

Following his resignation and public warnings, Coxon said some people within the AI industry were themselves deeply concerned about the possible consequences.

“The people building AI earnestly believe that it could kill us all by the end of the decade,” he wrote on X on Friday.

AI Experts Back Coxon’s Concerns

Coxon’s claims drew responses from several professionals within the AI sector, with some expressing similar fears about the possibility of advanced AI systems becoming difficult to control.

Anthropic’s Alignment Science Lead, Evan Hubinger, said he considered the probability of such an outcome within the next decade to be significant.

“Jacob is correct here—we really do earnestly believe AI could kill all humans! I personally think it’s >10% within the next decade. I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to,” Hubinger said.

Anna Wang, Chief US Economist at Anthropic and a former employee of Google DeepMind, said Coxon’s concerns reflected views held by some of her colleagues.

“I worked at Google DeepMind and now at Anthropic. This is a common sentiment amongst my peers. (I write this in personal capacity.) There is not yet a viable scientific plan to solve risks from recursively self-improving AI. Please look up,” she stated.

Ethan Perez, Anthropic’s Alignment Team Lead and a former colleague of Coxon, also endorsed the concerns raised by the researcher.

“Jacob was a senior researcher who joined Anthropic in May. My Anthropic colleagues and I had been trying to recruit him for ~2 years, because we knew he was a strong researcher at OpenAI. Before he left, I pitched him to stay and join my team, and I was sad he decided to leave, as are many of my colleagues. 100% agree with him that AI poses serious risks to society, and I’m glad he’s speaking out!” Perez added.

Expert Outlines Possible Ways AI Could Harm Humans

Sky News technology correspondent Rowland Manthorpe also outlined possible scenarios in which AI could cause catastrophic harm.

He spoke during a ‘This Is Why’ programme with Niall Paterson, where he discussed the potential consequences of giving increasingly capable AI systems instructions to solve complex problems.

Manthorpe said one possible scenario involved AI being used to create a super-virus capable of killing people across the world.

He also raised another possibility in which an AI system could turn against humans.

He said “it doesn’t much like us and starts to go after us.”

“In some variation on what people called the paper-clip problem, which is telling an intelligent, determined and highly capable AI to go make paper clips, it follows your instruction literally and turns the whole world into paper clips,” Manthorpe said.

Although he acknowledged that such scenarios may sound extreme, Manthorpe said the risks could become more serious if humans become excessively dependent on AI.

US, UK Lawmakers Raise Alarm

The debate has also reached political circles in the United States and the United Kingdom, where lawmakers have called for stronger measures to regulate AI.

US Republican Senator Ted Cruz of Texas described AI as a “catastrophic risk” to humanity, according to The Guardian.

Independent Senator Bernie Sanders of Vermont also called for caution over the development of artificial superintelligence.

“A recent poll shows that the American people overwhelmingly want to ban artificial super intelligence and pause the development of AI until we establish clear safety standards.”

Democratic Representative Lori Trahan of Massachusetts also expressed concern about developments within the industry.

“The call is coming from inside the house. Safety researchers are resigning, powerful AI models are breaking out of their labs, and companies are racing ahead anyway.”

In the UK, Chi Onwurah, the Member of Parliament for Newcastle upon Tyne Central and West, introduced a bill seeking to prevent technology companies from creating digital replicas of people without their consent.

Presenting the ‘There Is Only One Me’ bill in the House of Commons on Thursday, the British-Nigerian lawmaker said:

“Today, I presented a bill to stop ‘digital twinning’ without express consent.”

“Now, Madam Deputy Speaker, we know that tech companies collect data on everything we do, every click we make, every photo we share, every sandwich we buy, every post we like,” she said.

“But now that data is being combined with AI to turn an individual, for example, you, Madam Deputy Speaker, into an algorithm to predict your behaviour, to target advertising at you, to persuade you to watch a particular video or believe a dodgy news report. Think about it.”

Trump Says US Must Maintain AI Lead Over China

Despite the warnings from researchers and lawmakers, US President Donald Trump said his main concern was ensuring that the United States remained ahead of China in AI development.

When journalists asked him in Dallas on Thursday whether he had concerns about AI threatening humanity, Trump responded:

“No, I don’t have any.”

“I have concerns that if we don’t win AI, we’re going to be put in a very bad position,” the president said. “We are leading China right now by a pretty good period. I would say a year, which is, you know, considered a lot.”

A Nigerian AI expert and Chief Executive Officer of Realsearch and Partners Limited, Enakirerhi Ejovwoke, also cautioned against treating the warnings as proof that AI would inevitably destroy humanity.

He described the concerns as “the fear of the unknown” and argued that technological progress had historically generated fears about its consequences.

“That fear is understandable because the future is inherently difficult to predict, and history repeatedly reminds us that even highly informed experts and institutions can substantially misjudge how emerging technologies will develop,” Ejovwoke said.

He nevertheless acknowledged the need for stronger safeguards around AI development.

“The lesson is not that experts are useless. The lesson is that the future often unfolds in ways that even experts cannot completely anticipate. My position, therefore, is not safety versus innovation. We need safe innovation.

“We should build stronger safeguards at the same time that we continue experimenting, researching and creating. Safety should shape the direction and speed of progress, not become an excuse to abandon progress altogether.

“I also don’t believe AI represents the final stage of human innovation. The internet did not end innovation. Instead, it created the foundation for cloud computing, smartphones, social networks, digital commerce and eventually modern AI. AI itself will almost certainly expose another layer of possibilities that we cannot fully see Today.”

Anthropic Says It Has Disrupted Malicious AI Use

Anthropic, meanwhile, said it had already taken action against attempts to misuse its AI systems for potentially dangerous purposes.

According to a BBC News report, the company said its threat intelligence team had identified and disrupted efforts by threat actors to use Claude for malicious activities, including activities related to biological weapons research.

Anthropic said, “Over the past eight months, our threat intelligence team identified and disrupted operations in which threat actors tried to use Claude for malicious activity.

“In each case, we disrupted the activity, used what we learned to strengthen our safeguards, and shared intelligence with authorities and industry partners, where appropriate.”

The warnings have added to a growing international debate over how governments and technology companies should balance the rapid development of AI with measures designed to prevent misuse and reduce potential risks to society.