Softbank's Masayoshi Son has invested billions in AI. Now he says it could be ‘super dangerous’

SoftBank Group chief Masayoshi Son, one of the biggest financial backers of the artificial intelligence boom, has issued an unusually stark warning about the technology he has spent years championing.

As AI systems become more capable at an accelerating pace, Son said countries must learn to trust one another and work together because advanced AI in the wrong hands could become “super dangerous”.

Speaking at an event on the sidelines of the Science and Technology in Society forum in Kyoto on Sunday, Son said the rapid expansion of AI capabilities was making international cooperation increasingly urgent. He appeared alongside Michael Kratsios, the White House science and technology policy adviser, who was in Japan to promote a US-led initiative on using AI to accelerate scientific research.

“We don’t have any more luxury of having human fight against human,” Son said, arguing that countries need to cooperate to manage the much larger threat posed by increasingly powerful AI systems.

The comments offer a striking contrast with Son’s long-standing enthusiasm for AI. The 69-year-old Japanese billionaire has directed nearly $65 billion of SoftBank funding towards OpenAI, the creator of ChatGPT, and has repeatedly argued that AI and intelligent machines could transform the global economy.

Masayoshi Son’s AI warning comes as machines gain new capabilities

Son said AI services are continuing to improve at an exponential rate. The technology that initially powered chatbots is increasingly being used to browse websites, write software and perform more complex tasks, while consuming growing amounts of general-purpose computing power.

He expects the technology to enter another stage next year, when AI systems will increasingly be able to help robots and other machines interact with the physical world.

That transition could significantly widen the consequences of AI development. Systems that can reason, code and navigate digital environments are already capable of carrying out increasingly complex tasks. Giving similar capabilities to machines operating in the physical world could extend AI's reach well beyond computers and smartphones.

Son warned that the stakes will depend not only on how capable these systems become, but also on who controls them.

“Superintelligence in the wrong hands can become ‘super dangerous,’” he said.

US leads 17-country push to use AI for scientific research

Kratsios was in Kyoto to announce a US-backed initiative involving 17 countries aimed at using AI to speed up scientific discovery.

Representatives from countries including Japan, South Korea, the UK, Germany, Singapore and the United Arab Emirates joined Kratsios on stage. China was not among the countries that initially endorsed the White House statement.

“In automated cloud laboratories operating at industrial scale, robotic systems will conduct hundreds of parallel experiments, generating data that feeds back for real-time analysis and hypothesis refinement,” Kratsios said.

He said machines can also process vast amounts of scientific literature and compress years of research into “tens of hours of work by armies of digital agents assisting human researchers.”

The promise is substantial. AI could allow scientists to test more hypotheses, analyse larger datasets and automate experiments that would otherwise take years.

But the same acceleration is also at the centre of Masayoshi Son’s warning.

SoftBank’s Masayoshi Son sees AI and machines reshaping the global economy

Despite his concerns about safety, Son remains one of the technology sector’s most vocal believers in the economic potential of AI.

He has maintained his long-term prediction that AI and learning-capable machines could account for at least 20% of global gross domestic product by 2040. At today’s scale of the world economy, that would amount to about $46 trillion.

For SoftBank, that expectation has translated into a huge financial commitment to AI. Son has positioned the Japanese conglomerate around what he sees as a coming wave in which increasingly intelligent software, chips, robotics and computing infrastructure become tightly connected.

His latest comments suggest that he sees international coordination as increasingly important as that transformation gathers pace.

“This is just the beginning,” Son said.

(With inputs from Bloomberg)