High Likelihood of AI Correction

A market correction in the artificial intelligence sector is “pretty high,” according to Nick Clegg, former Meta executive and British politician. Clegg, speaking on Wednesday, tempered the enthusiasm surrounding artificial superintelligence, suggesting current valuations may be unsustainable.  

Clegg, who previously served as the U.K.’s deputy prime minister before shaping policy at Meta, observed that the AI boom has fueled “unbelievable, crazy valuations.”  

“There’s just absolute spasm of almost daily, hourly, deal making,” he told CNBC, emphasizing the rapid pace of investment.  

“You’ve got to think, wow, this could be headed for a correction,” he stated, assessing the likelihood of a correction as “pretty high.” The concern stems from whether current valuations accurately reflect the underlying potential or constitute a speculative bubble.

Market bubbles are typically characterized by valuations that detach from fundamental business performance, whether in private or public markets.

Clegg believes a key factor in any potential correction lies in the ability of large hyperscalers – those “pouring hundreds of billions of dollars into the ground and building these data centers” – to realize a return on their substantial infrastructure investments and establish sustainable business models. This hinges not only on technological progress but also on market adoption and the monetization of AI-powered solutions.

“That’s obviously going to raise some issues,” he added, pointing to the “fundamental paradigm on which this whole industry is built, the so-called large language model AI paradigm.” The dependence on large language models, while currently dominant, could prove limiting, especially if more efficient or alternative approaches emerge.

The debate revolves around ambition: striving for artificial superintelligence, defined as AI surpassing human intelligence – often hailed as the ‘holy grail’ – versus focusing on artificial general intelligence, where AI systems achieve human-level capabilities. 

While figures like SoftBank founder Masayoshi Son and Meta CEO Mark Zuckerberg are betting on artificial superintelligence, even establishing dedicated AI labs, Clegg remains skeptical. 

“I think there are certain limits to that probabilistic AI technology, which means that it won’t perhaps be quite as all singing and all dancing as people suggest,” Clegg added. However, he acknowledged the enduring potential of AI: “But it doesn’t mean that technology itself is not going to persist, it’s not going to flourish and is not going to have a huge effect.”

Clegg’s position highlights a contrasting perspective on the short-term and long-term potential of AI. While acknowledging the genuine advancements and transformative possibilities, he advocates for a more measured approach to hype and expectations.

Historical parallels offer some context. Meta, after surviving the dot-com bubble, become a tech titan. Amazon and Google followed similar paths. This illustrates that bubble bursts don’t necessarily spell doom.

Venture capitalists often emphasize that the strongest companies emerge from economic downturns or constrained funding environments. This is because investors place a premium on fiscal discipline and strong business metrics, compelling leaders to operate more efficiently and innovate with limited resources.  

Clegg’s reservations echo concerns voiced by other investors and tech leaders. They share the view that a bubble may be forming, without discounting AI’s staying power. 

The influx of investment signals an “industrial bubble,” according to Jeff Bezos, while underscoring the certainty that “AI is real, and it is going to change every industry.” 

Clegg anticipates a staggered adoption of AI, prioritizing easily integrated applications. 

“There’s a lot of hype. People in Silicon Valley assume that if you invent a technology on Tuesday, everybody’s going to use it on Thursday. It’s not actually how it works at all,” he said.

“It took 20 years for all of us to get onto desktop computing after desktop computing was technologically feasible. So, I think it’s the pace that is the thing to look out for. That will vary sector from sector to sector, country by country, but I think it might be just a little bit slower than some of the technologists themselves are predicting at the moment,” he added.  The key, according to Clegg, lies in managing expectations and acknowledging the time required for widespread adoption and integration across various industries.