AI extinction warning sparks global debate over trust, regulation

A recent post on X turned artificial intelligence from a technical debate into an international public anxiety. Jacob Coxon, a researcher who had worked on pretraining at OpenAI and Anthropic, announced that he was resigning from Anthropic.
He accused both companies of racing toward self-improving superintelligence and 'gambling with our lives.' The wider thread warned that advanced AI could kill everyone by the end of the decade. His post received more than 160 million views, spreading far beyond the United States and Europe into the Arabian Gulf, including Oman.
Some details circulating online require correction. Coxon did not claim that he had worked in AI for only a few months. He wrote that he had spent three years conducting pretraining research across OpenAI and Anthropic, although his time at Anthropic appears to have been much shorter than his total industry experience. The claim that humanity will be destroyed by 2030 is also a warning, not an established prediction. Coxon raised a serious possibility, but he did not provide evidence proving that such an outcome will occur.
The post's extraordinary reach created a second controversy. People were no longer discussing only whether AI might become dangerous. They were also asking why a relatively unknown account had suddenly appeared in feeds around the world.
Elon Musk reacted skeptically, describing the episode as seeming like a setup and later characterising the surrounding campaign as a 'psyop.' His reaction intensified suspicions that the post had been deliberately amplified to frighten the public and prepare it to accept stricter controls over AI and social media.
That suspicion is understandable, but it should not be presented as fact. There is currently no publicly demonstrated proof that X, a government, Anthropic, or another organisation artificially pushed the post.
Social-media virality can result from recommendation systems, reposts by influential accounts, news coverage, and the emotional power of a frightening message.
A warning from someone with experience inside leading AI laboratories contains all the ingredients for explosive attention. The unusual spread deserves scrutiny, but asking who benefited from it is different from proving that someone engineered it.
The debate should therefore avoid two extremes. One extreme treats human extinction as practically guaranteed and encourages panic. The other dismisses every warning as propaganda and assumes that technological progress will automatically benefit humanity.
A wiser position recognises genuine risks without surrendering to fear. Advanced AI may create dangers involving autonomous weapons, cyberattacks, misinformation, unemployment, surveillance, and the concentration of power. These problems justify public oversight, independent safety research, and transparent rules. However, regulation should protect citizens rather than use fear as an excuse to remove their agency or silence legitimate discussion.
This controversy also raises a deeper question: What do we want AI to do for us? Many people would prefer machines to handle exhausting, repetitive work - cleaning homes, washing dishes, or performing hazardous industrial tasks - while humans retain meaningful and creative activities. Instead, AI is often being developed first to write books, produce images, imitate artists, and make films. This feels backward. Technology should free people from drudgery, not make human imagination seem disposable.
The viral warning matters, but not because it proves that everyone will die by 2030. It matters because it reveals a worldwide crisis of trust: distrust of AI companies, algorithms, governments, billionaires, and even viral information itself. The responsible response is neither panic nor denial. It is to demand evidence, preserve human choice, regulate proportionately, and direct AI toward work that genuinely improves human life.

DR NAJMA AL ZIDJALY The writer is an Associate Professor, Sultan Qaboos University