SoftBank CEO warns of AI safety risks, urges global cooperation

SoftBank Group Corp. chief Masayoshi Son, a leading AI proponent, warned that the exponential rise in artificial-intelligence abilities makes coordinated safety efforts essential. Speaking at the Science and Technology in Society forum in Kyoto alongside White House science and technology policy adviser Michael Kratsios, Son said nations must develop trust to jointly manage the threat posed by powerful AI models, which could become "super dangerous" if misused.

He highlighted recent security breaches in AI systems as evidence of growing risk. Son, who has directed nearly $65 billion into OpenAI, also projected that AI and machine learning will account for at least 20 % of global GDP, roughly $46 trillion, by 2040. Kratsios announced a US-led coalition of 17 countries aiming to accelerate scientific research through AI, with participants including Japan, South Korea, the UK, Germany, Singapore and the United Arab Emirates. Son added that AI services will soon move beyond chat and code generation to support robots and other machines interacting with the physical world as early as next year.