AI apocalypse now?

Even more cynically, these companies may be hyping up the danger ahead of releasing their own cybersecurity products, thereby inflating a problem that only they can solve.

AI’s capacity to destroy humanity in a spectacular fashion in the future obscures the very real harms it is already causing now

New Internationalist, 5 November 2026
Decca Muldowney turns a critical eye on industry-theatrics to examine the real and imagined world-destroying potential of AI.

In early September Jacob Coxon, a researcher with the AI company Anthropic, very publicly quit his job, sharing a dire warning about the supposed existential risks of the technology.

‘The people building AI earnestly believe that it could kill us all by the end of the decade,’ he wrote on X. ‘No other human activity poses this level of danger.’ He blasted both Anthropic and its competitor, OpenAI, for not doing enough to prevent harm by AI.

Coxon was not the first to make such an extreme claim. Many in the cult-like inward-looking world of Silicon Valley have been expressing similar predictions of an AI-induced apocalypse in the last few years. Some of these ‘doomers’, like Geoffrey Hinton, the so-called ‘godfather of AI’, have made entire careers of it.

But this time the message cut through to the general public, sparking a debate across the world about the potential dangers of AI and whether it truly has the ability to kill us all. Shortly after Coxon’s pronouncement, Dario Amodei, the chief executive of Anthropic, published an essay calling for a global slowdown of AI development. Sam Altman, of OpenAI, followed suit. But why are the people building this technology calling for it to be better regulated? Why, if they truly believe it has world-ending capabilities, are they continuing to develop it? Why do they want to keep the public afraid?

Some believe these companies are engaged in a form of regulatory capture, by which private companies aim to direct and control the way regulation is applied to their industry, thereby foreclosing possibilities for companies that come after them while maximizing their own benefit. After all, despite Amodei’s warning, his company is moving ahead with its initial public offering (IPO), meaning the company will go public. Anthropic has a potential value of $2 trillion.

Even more cynically, these companies may be hyping up the danger ahead of releasing their own cybersecurity products, thereby inflating a problem that only they can solve

AI’s capacity to destroy humanity in a spectacular fashion in the future obscures the very real harms it is already causing now.

The Apocalypse ploy

The fears of AI ending our world – whether by turning on its creators or unleashing bioweapons – seem at first to be in direct contrast with the claims of AI ‘boosters’ that believe the technology will change our world for the better.

But a more helpful lens through which to understand this is ‘critihype’, coined by Virginia Tech professor Lee Vinsel, which describes negative discourse about technology that ‘retain[s] the picture of extraordinary change but focus[es] instead on negative problems and risks’.

But doomers and boomers can also be seen, counterintuitively, as ‘two sides of the same coin’, as tech critics Emily M Bender and Alex Hanna write in their book The AI Con: ‘the substance of the coin is the belief that the development of AI is inevitable and that resulting technology will be both autonomous and powerful, and ultimately beneficial, if we play our cards right. It’s only the faces of the coin that differ’.

What is clear is that the focus on AI’s capacity to destroy humanity in a spectacular and dramatic fashion in the future obscures the very real harms it is already causing now. Among the most pressing of these is its contribution to the climate crisis, a real and present danger that does threaten the entire planet – or at least its capacity to support life as we know it.

Another arena in which AI is already causing distinct and severe harm is in modern warfare, particularly the determination of arms companies and governments to produce autonomous weapons.

Unmanned drones have already become normalized but now AI is being incorporated into targeting systems. The genocide in Gaza was proudly described in the Israeli press as the ‘first AI war’. The Israel Defence Forces (IDF) used AI systems to locate targets in Gaza in real-time, killing thousands. The use of Palantir-built AI technology by the US-Israeli forces contributed to the killing of 150 school children in Minab, Iran. Here there is no need for grand stories about the end of the world – these children’s worlds were ended in a moment.

Techlash 2.0

There are some rays of light amid the doom. In the United States and beyond, previously pro-AI politicians have been pivoting after huge amounts of public backlash to the construction of data centres, which are proving to be deeply unpopular across the political spectrum. Between April and June this year, at least 45 data centre projects were delayed or blocked after local opposition, at a cost of $68 billion to the industry.

Similar fights against data centre construction are taking place in Uruguay, the Netherlands, Chile and beyond. Also in the US, fears around surveillance have solidified into a growing campaign against Flock cameras, AI-powered automatic license plate readers that are now widespread.

The AI doomers would like us to believe their apocalyptic visions, as long as we keep ignoring their present-day harms. They would like us to stay afraid, so that we do not organize. We should not allow them to dictate the terms of humanity’s future. https://newint.org/science-and-technology/2026/ai-apocalypse-now?utm_source=ni-email-whatcounts%20&utm_medium=email&utm_campaign=2026-11-06%20enews%20-%206%20november%20enews&utm_content=1%20NI%20Main%20List1%20-%20enews%20-%20International%20AND%20North%20America