Daily Times Editorials 10th September 2026

Playing With Our Lives

Anthropic insider Jacob Coxon has walked away from the AI industry with an accusation that deserves more than Silicon Valley’s usual shrug. After working on training frontier models at both OpenAI and Anthropic, the 27-year-old researcher says the leading artificial-intelligence laboratories are racing toward self-improving superintelligence and “gambling with our lives.”

While technologists have been predicting utopia and apocalypse in roughly equal measure for decades, it is disturbing to see where the latest warning comes from.

Coxon was not watching the AI boom from the outside. He was helping build it. Nor does he accuse Anthropic of being oblivious to the danger. Quite the opposite. His case is that researchers inside the company understand how badly things could go and continue racing because stopping unilaterally might allow somebody less cautious to win.

There is something profoundly unsettling about that logic. Everybody may know the cliff is there yet nobody wants to take a foot off the accelerator.

Another Anthropic researcher Evan Hubinger has publicly put his own estimate of AI causing human extinction within the next decade above 10 per cent. Let that sink in. A scientist working near the centre of one of the world’s most advanced AI laboratories believes the probability of the technology killing everyone may be greater than one in ten-and development continues.

AI enjoys a strange privilege. The companies creating the risk also conduct much of the testing, determine many of the safeguards and decide when another model is ready to leave the laboratory. And to add to the worries, its capabilities are no longer confined to chatbots producing embarrassing answers. AI agents can write software, operate digital tools, conduct research and perform increasingly long sequences of actions with diminishing human supervision. Developers have already encountered behaviour they did not intend and have responded with stronger monitoring, containment and shutdown mechanisms.

None of this proves that a machine is about to seize control of civilisation. Still, it establishes something uncomfortable enough: we are building systems whose behaviour can surprise their makers while simultaneously trying to make those systems vastly more capable.

Washington is beginning to notice. Proposals for a federal AI regulator, mandatory safety standards and even restrictions on superintelligence are moving from the fringes towards Congress. Whatever form regulation eventually takes, the present arrangement cannot survive the technology it is supposed to govern.

The darkest possibility is not that reckless men are knowingly building a machine that could destroy humanity. It is considerably more plausible, and therefore more frightening: intelligent people may understand the danger, believe somebody will build it anyway, and decide that it had better be them. *

Great Test

The Houthi barrage on southern Saudi Arabia has emerged as an operational test for the Makkah Joint Defence Agreement barely a month after it was signed. Tuesday’s missile and drone attacks hit Aramco facilities and an airbase across four Saudi cities, injured 73 people and pushed a Yemen conflict that Riyadh had spent years trying to contain back across the border. Defence Minister Khawaja Asif has now said that further spillover could bring the Pakistan-Saudi-Türkiye pact into operation.

What that means needs defining before another round of missiles decides it for the three capitals. The agreement signed on August 7 commits its members to regard an armed attack on one as an attack on all. Yet its public text does not spell out what assistance follows, against whom or at what threshold. This was always going to matter in a region where wars are fought not only between states but through armed movements with varying degrees of external sponsorship.

The Houthis illustrate the difficulty. They are closely aligned with Iran and have benefited from Iranian military support, but Tehran’s relationship with them does not make every Houthi operation an Iranian attack. Their latest strikes also came amid renewed Saudi-Houthi fighting in Yemen. Collapsing those distinctions would turn a defensive arrangement into something considerably more dangerous.Browse Islamic Supplies

Pakistan has a useful precedent of its own. In 2015, parliament refused to join the Saudi-led intervention in Yemen while simultaneously pledging to stand by the Kingdom if its territorial integrity were threatened. That distinction was neither equivocation nor abandonment. It kept Pakistan out of a ruinous war without pretending that Saudi security was of no concern to Islamabad. Makkah Pact should not erase that doctrine. Rather, it should give it even greater precision.

Pakistan already has a substantial defence presence in Saudi Arabia, with around 8,000 Pakistani personnel, along with aircraft, drones and air-defence capabilities deployed there. There is consequently a wide field for cooperation, may it be intelligence, missile and drone defence, early warning, protection of critical installations and maritime security, which does not require Pakistani forces to become combatants inside Yemen.

Nor can Iran be wished out of the equation. Pakistan shares a long and occasionally troubled border with it. Cross-border militancy, Afghanistan and trade make a workable relationship with Tehran a permanent national requirement. Islamabad’s utility lies precisely in not being reducible to one regional camp.Learn Arabic Online

The wider setting makes restraint harder, but all the more necessary. The US-Iran conflict has already impaired energy traffic through Hormuz and the Houthi attacks have now put Saudi oil infrastructure under renewed pressure from the south. Brent has crossed $100 amid the latest maritime escalation. For Pakistan, every prolonged Gulf crisis eventually reappears in fuel prices, electricity tariffs and inflation. This also explains why the Makkah Pact matters beyond its military clauses. Riyadh, Ankara and Islamabad are searching for greater regional ownership of security at a time when confidence in the old, overwhelmingly American security architecture has weakened. But regional ownership should mean greater capacity to contain wars, not merely greater capacity to join them. *