AI poses ‘Hiroshima’-style threat to humanity without global rules, says Cooper ↗
UK foreign secretary Yvette Cooper warned that governments cannot wait for a catastrophic AI incident before agreeing international rules. She wants major powers, particularly the US and China, to work together on common guardrails.
Her concerns extend beyond runaway systems. State-backed criminals, extremists and hostile organisations could use AI to undermine security and democracy - a rather grim toolbox, by any measure.
Trump's AI plan has governments and tech firms panicking ↗
America's influence over the most powerful AI models is becoming a sovereignty problem. Governments and companies increasingly fear that access to frontier systems could be restricted, delayed or wielded as political leverage.
Buying an American AI service is not the same as controlling critical infrastructure... uncomfortably, it is closer to renting the engine of your economy from someone who keeps the keys.
Why OpenAI and Anthropic may struggle to float ↗
Private investors have pushed OpenAI and Anthropic towards extraordinary valuations, but public markets may demand something less sparkly - sustainable cash flow. Frontier research remains brutally expensive, while cheaper models are making the technology increasingly interchangeable.
Infrastructure and hosting are capturing far more generative-AI revenue than foundation-model developers. Add competition from Microsoft, Google and Amazon, and those blockbuster listings suddenly look complicated... or so it seems.
Big Tech Has Suddenly Flipped on the AI Jobs Wipeout Scenario ↗
Tech leaders who once openly predicted enormous job losses are adopting a noticeably sunnier message. They now emphasise human-AI collaboration, productivity and keeping workers central to the economy.
The systems may have evolved, sure - but public opinion towards AI has also deteriorated. Perhaps the technology changed, perhaps the sales pitch did... probably both.
UK regulator warns of ‘arms race’ to keep up with AI use in financial services ↗
The Financial Conduct Authority says regulators must adopt AI themselves to monitor how rapidly banks, fintech companies and consumers are using it. Millions are already consulting chatbots about savings, borrowing and other financial decisions.
AI could personalise financial services, but it could also enable biased recommendations, opaque pricing, fraud and unnervingly precise manipulation. The watchdog wants stronger powers and a review of AI services operating outside normal financial regulation.
SK Telecom Pursues 15GW AI Data Center Buildout, Aiming to Become Asia’s AI Infrastructure Hub ↗
SK Telecom unveiled plans for an enormous AI data-centre network, beginning with 5GW of capacity and ultimately targeting 15GW. The company wants South Korea positioned as a major Asian base for AI computing.
SK Group plans to combine energy, semiconductors, connectivity and data-centre operations into one full-stack infrastructure push. It is wildly ambitious - essentially industrial policy wearing a server-rack costume.
FAQ
Why are global AI rules becoming more urgent?
Governments fear that waiting for a major AI-related disaster could mean reacting only after the damage is done. International rules could establish shared safety standards, clarify responsibilities and reduce the risk of powerful systems being misused. Cooperation between major AI powers, particularly the US and China, is widely seen as essential, since national regulations may not be enough to address threats that cross borders.
What security risks could stronger AI regulation address?
AI regulation could help governments respond to threats from state-backed criminals, extremist groups and hostile organisations. These actors may use AI to spread disinformation, automate cyberattacks or interfere with democratic processes. Common guardrails would not remove these dangers entirely, but they could strengthen oversight, information sharing and accountability between countries.
Why are governments worried about relying on American AI companies?
Access to advanced AI models is increasingly being treated as a matter of sovereignty. Governments and businesses may depend on systems controlled by US companies, which means access could potentially be restricted, delayed or shaped by political decisions. A common response is to invest in domestic infrastructure, diversify suppliers and retain greater control over critical data and computing capacity.
Why might OpenAI and Anthropic struggle with public listings?
Private valuations can reflect expectations of rapid future growth, while public investors tend to scrutinise revenue, costs and sustainable cash flow more closely. Frontier AI research and computing infrastructure remain extremely expensive. Competition from cheaper models and major technology companies could also make foundation-model services appear more interchangeable, putting pressure on margins and long-term valuations.
How could AI affect jobs despite more positive messages from tech leaders?
Technology executives increasingly describe AI as a tool for collaboration and productivity rather than a direct replacement for workers. In practice, the effects will probably differ by role, industry and workflow. Some tasks may be automated, while others may be redesigned around human oversight, judgement and creativity. This makes retraining and responsible deployment increasingly important.
How is AI regulation affecting finance and data-centre investment?
Financial regulators are considering stronger powers as consumers use chatbots to make decisions involving savings, borrowing and other services. Concerns include biased recommendations, opaque pricing, fraud and highly targeted manipulation. At the same time, companies such as SK Telecom are planning vast data-centre networks, showing that AI governance and physical infrastructure investment are moving forward together.