AI leaders call for a slower pace
Dario Amodei, chief executive of Anthropic, has warned that the rapid development of artificial intelligence could create systems that are difficult to control. In a recent essay, he urged the industry to adopt a more deliberate approach.
"We need to ensure alignment before we push the frontier further," Amodei wrote.
His call was echoed by Sam Altman, chief executive of OpenAI, who told EuroHerald's editor-in-chief that an initial public offering would be ill-advised given current safety concerns.
"As we have said, we're not rushing into an IPO," Altman said. "I actually think that, given everything happening with safety, right now would be an ill-advised moment to go public."
Other prominent figures added their voices. Elon Musk simply wrote, "Dario is right." Satya Nadella, chief executive of Microsoft, welcomed research that prioritises alignment, while Demis Hassabis of Google DeepMind praised Amodei's proposal for an industry-wide standards body.
Former White House AI adviser David Sacks suggested that the onus lies with the companies themselves, urging them to lead the effort. "Go ahead. You guys are the frontier," he wrote.
Even political leaders weighed in. Mike Johnson, speaker of the U.S. House of Representatives, said a moratorium was unrealistic given competition with China, and former President Donald Trump dismissed proposed guardrails as ineffective.
Chinese labs gain efficiency
While the debate over speed continues, analysts note that Chinese AI laboratories are closing the performance gap with U.S. firms by improving computational efficiency. According to Brendan Burke of Futurum Group, Chinese researchers have devised algorithms that reduce the complexity of attention calculations by an order of magnitude, allowing them to summarise the most relevant tokens more effectively.
Burke explained that limited compute resources forced Chinese labs to innovate rather than simply increase hardware power.
Ameya Kanitkar, co-founder of the AI measurement platform Larridin, quantified the impact: Chinese models handle roughly 75% of engineering tasks "reasonably well" at one-fifth the cost of comparable U.S. models.
Retailers see AI as a human-centred tool
In a separate development, Malina Ngai, group chief executive of AS Watson, argued that artificial intelligence should augment, not replace, staff. Speaking at the EuroHerald Leaders Forum, she questioned whether AI would cut 30% of the workforce or enhance employee intelligence.
"Is it really that, with AI, we're going to cut out 30% of our workforce? Or… are we going to increase the quality of our people's intelligence?" Ngai asked.
Ngai described a "human-AI partnership" strategy that has boosted employee engagement and improved a "customer love score" as staff spend more time interacting with shoppers.
Other tech headlines
- Matt Mullenweg survives an ouster at Automattic.
- SoftBank secures a nearly $12 billion loan to support its OpenAI investment.
- The U.S. Interior Secretary meets AI hyperscalers to discuss data centre use on federal lands.
- Bending Spoons raises prices across its portfolio.
- The NSA plans a restructuring into five divisions focused on AI, China, cybersecurity, warfighting and global intelligence.
- Nvidia is reported to be involved in Anthropic's potential $10 billion IPO.
- Amazon workers on food stamps have tripled despite record revenue.
What lies ahead?
The call for a slower, safer AI trajectory faces a complex landscape. Industry leaders are urging self-regulation, yet geopolitical competition and rapid technological advances, especially from China, may limit the effectiveness of voluntary pauses. Observers suggest that without coordinated standards, the sector could fragment, with divergent safety practices emerging across regions.
For European policymakers, the debate underscores the need to balance innovation with robust oversight, potentially shaping future EU AI regulations.