Will GenAI deliver real transformation?

 IS THE ARTIFICIAL Intelligence (AI) bubble about to burst? Not yet, not quite. Is it a bubble at all? Generative Artificial Intelligence (GenAI) has powerful advocates. Elon Musk, the world’s richest man, said in a recent podcast that “superintelligence” will run everything by 2030.

What Musk means by superintelligence is a GenAI module trained to be not only more intelligent than any living human but possessing both reasoning power and empathy—two qualities sceptics say GenAI today lacks.

Meta CEO Mark Zuckerberg is even more bullish on AI than Musk. He has created a division within Meta called superintelligence and hired million-dollar AI researchers to work on advanced GenAI models. He is pouring over $100 billion in building giant datacentres that house supercomputers on which GenAI models can be trained. These datacentres are gluttons for power: many need over 5 gigawatts (GW), enough to power a US city of two million people.

The first warning signs that exuberance over a GenAI-run superintelligent future was overdone came last week. The world’s most aggressive venture capital firm, Softbank, sold its entire shareholding of $5.83 billion in AI chipmaker Nvidia, the world’s most valuable company with a market capitalisation of $5 trillion.

Softbank Founder and CEO Masayoshi Son said the sale was to raise funds for investing in other stocks, including several AI companies. The Japanese venture capitalist already has stakes in AI firms, OpenAI and PerplexityAI, so Son’s explanation wore thin. Softbank’s CFO Yoshimitsu Goto was circumspect: “I can’t say if we’re in an AI bubble or not.”

For one of the world’s largest venture capitalist firms with a net profit in the July- September quarter of $16.6 billion (`1.50 lakh crore) and a market capitalisation of $210 billion (`18.61 lakh crore), Softbank’s explanation for the Nvidia sale raised eyebrows. GenAI is unquestionably a transformative technology. But the narrative built around it by firms like Zuckerberg’s Meta, Sam Altman’s OpenAI and Musk’s Xai is partly aimed at attracting investor funding for AI “use cases” that so far have been underwhelming. Altman has himself often used the word bubble to describe the GenAI boom, calling some company valuations “insane”.

The debate on GenAI has now turned to whether the technology can overtime pose an existential danger to humankind. Some of the alarm is overheated and plain wrong. But the creation of “superintelligence” has drawn warnings from Geoffrey Hinton, the 2024 Nobel Prize winner for physics who is regarded as one of the godfathers of AI.

Hinton told BBC Radio 4 after winning the physics Nobel last year that there was a “10 per cent to 20 per cent chance” that AI would lead to human extinction within the next three decades. That is clearly a case of super-anxiety over the imagined destructive power of superintelligence. Hinton, however, backed his views with action. He resigned last year from Google which is at the vanguard of GenAI development.

Hinton added: “My worry is that just leaving it to the profit motive of large companies is not going to be sufficient to make sure they develop it safely. Most of the experts in the field think that probably within the next 20 years, we’re going to develop AIs that are smarter than people. And that’s a very scary thought.”

It isn’t actually. The other godfather of AI, Yann LeCun, chief scientist at Zuckerberg’s Meta, feels Hinton worries too much. He says: “AI could actually save humanity from extinction.”

As always, the truth lies between the two extremes. The real reason for the drama orchestrated by Altman, Zuckerberg and Musk over the superpower of GenAI is to attract investors to fund their companies’ expensive plans to create superintelligence. The huge datacentres being built across America need a tsunami of private equity funding. Without an existential narrative, they won’t get it.

But increasingly hard questions are being asked. What are the real-life use cases of GenAI? They have helped medical diagnostics, made coding almost instantaneous, and automated industry processes. So far, GenAI is about speed, not innovation. Humanoid robots in China strut around shopfloors, often colliding with each other and tumbling down. Many remember the excitement over Metaverse, IOT and Web3.0. Most have remained acronyms.

Will GenAI deliver real transformation? Masayoshi Son, for one, is hedging his bets.