Masayoshi Son warns AI superintelligence can turn "super dangerous"

Son's remarks came on the sidelines of the Science and Technology in Society forum in Kyoto, where he shared a stage with White House science and tech policy advisor Michael Kratsios. His message marked a shift in tone from an executive who has spent years arguing that AI would deliver investment returns and help humanity — and whose SoftBank funds have put nearly $65 billion into ChatGPT creator OpenAI.

The Kyoto appearance also tied the safety warning to concrete government action. Kratsios was in the Japanese city to announce a U.S.-led 17-country push to better use AI to accelerate scientific research, with representatives of Japan, South Korea, the U.K., Germany, Singapore and the United Arab Emirates joining him on stage. China was not among the countries that initially endorsed the White House statement.

Beyond the cautionary note, Son kept his long-term optimism intact, describing AI moving through phases — from chatbots to browsing sites and writing code, and next to helping robots and other machines interact with the physical world starting next year. He reiterated a projection that AI and machines with learning capabilities would account for at least 20% of global GDP by 2040, or the equivalent of $46 trillion. Whether his safety call translates into cooperation may hinge on whether the countries now aligning around AI for science can extend that trust to the far greater threat he described, a gap the article leaves open.