The AI Slowdown Debate: A Complex Road Ahead

When Experts Warn: A Shift in Tone

There has been no shortage of warnings from AI's top leaders about existential risks over the years. At the world's first AI safety summit in Bletchley Park in November 2023, the focus was on the most serious threats imaginable—posed by the most powerful AI models. Back then, many thought such concerns were the stuff of science fiction, with more mundane issues like job displacement and academic dishonesty taking center stage.

Fast forward three years, and those same voices are sounding increasingly uneasy. On Saturday, Dario Amodei, head of AI company Anthropic, urged a slowdown in development. This call was echoed by Sam Altman of OpenAI and Elon Musk—two prominent figures who have long been at the forefront of AI innovation.

"We need to slow down," said Amodei. "We are building systems that could potentially be dangerous, and we must take responsibility for that."

This growing chorus of concern is not just from tech insiders—it's also coming from employees who feel genuinely frightened about the future of humanity. A former Anthropic researcher told me, "Staff who were developing these systems are genuinely frightened for the future of humanity," underscoring how deep the unease has become.

Is a Slowdown Really Possible?

The idea of slowing AI development sounds like a straightforward solution to potential catastrophe. But in practice, it's proving to be anything but simple.

A fundamental challenge lies in enforcement. Who would police such a slowdown? Would we rely on the AI companies themselves to be transparent about their activities—and more importantly, their decisions not to act? That would require an unprecedented level of trust, something the tech sector has never quite earned.

Amodei's proposed three-point plan included independent monitoring of AI models during development, industry-wide regulation, and global regulation. Yet, critics remain skeptical about how these measures would actually be implemented. Ed Zitron, CEO of EZ Primary Research, pointedly asked, "Nobody has given a substantive explanation of what 'slowdown' means."

• Putting researchers from METR into every AI company?
• Democratic and global coordination?
• How would such an initiative be enforced?

Zitron also warned that halting training of AI models could leave companies vulnerable to competitors, especially in China. "Their margins might improve but their products will be captured in amber and distilled by Chinese labs," he said.

The Global Race: Can We Slow Down While Competing?

America's obsession with staying ahead of China in the AI race has been well-documented. President Trump recently claimed the US is ahead, stating, "Whoever wins AI, wins."

China isn't known for wanting to come second either. Even if a company pauses, others won't—and they'll simply overtake. It reminds me of the height of nuclear disarmament campaigns: everyone feared being left behind.

But even more unsettling is the question of how we'd coordinate this globally. The lack of clarity around what a 'slowdown' actually entails leaves us with a dangerous gap—how do we define, monitor, and implement it across borders and companies?

The UK's AI Journey: A Strategic Dilemma

In the UK, we've seen growing excitement about AI's potential in sectors like healthcare and economic growth. We're being encouraged to adopt AI more widely—both professionally and personally.

The industry is considered a key driver of national strategy. As one former government adviser put it: "There is no plan B." So, could a slowdown potentially derail that momentum?

Moreover, the AI industry is burning through enormous amounts of capital and resources, with little return on investment. Reports suggest several firms are disappointed with current results. Economists predict a potential market correction—a 'levelling'—as some companies struggle to survive.

If the AI boom continues unchecked, we could see some of today's giants become the most powerful mega-corporations in history. But with that power comes serious ethical and governance questions.

The Corporate Greed Factor

While existential concerns dominate headlines, I also hear another recurring theme: corporate greed. Sasha Luccioni, founder of Sustainable AI, said, "I'm not worried about the existential risks of AI, I'm worried about the corporate greed of the companies that are creating it."

This perspective highlights an important truth: many tech leaders may be more interested in profit than safety.

Professor Dame Wendy Hall, who has advised the UN on AI, draws a powerful analogy: "It's like a farmer having a bull in a field which escapes and causes destruction—and the farmer blames the bull." The issue isn't the technology itself but the lack of robust guardrails. As Hall puts it, "Of course it's not the bull's fault—it's the farmer."

Regulation: A Double-Edged Sword

Some believe that now is the time to step in with strict regulations—before it's too late. But others argue that over-regulating too soon could stifle innovation and even backfire.

Alexander Voica, from UK AI company Synthesia, said, "We know these systems are getting more powerful, but we don't know where and how they're going to be used, and we haven't figured out how to take full advantage of their potential. My only concern in rushing to regulate now is that it could actually backfire."

Adding another layer to this debate, some political figures have proposed a more sinister view: that the push for regulation is not about safety but control—"regulating AI into oblivion," as one conservative think tank put it.

What Does This Mean for the Future?

Whatever the case, the storm currently engulfing the AI industry may already have caused reputational damage to firms in the lead. As Professor Hall aptly asked, "Would you invest in a company that says it's going to bring about human extinction?"

The truth is, the decisions made today—whether in boardrooms or policy centers—will shape not only how AI evolves but whether it serves humanity well. And with so many unknowns, the path forward remains unclear.

For now, we're left grappling with one fundamental question: How do we balance innovation with responsibility when the stakes couldn't be higher?