Masayoshi Son: Superintelligence in Wrong Hands Could Be 'Extremely Dangerous'

SoftBank founder Masayoshi Son issued a rare warning on Sunday at the Science and Technology in Society (STS) Forum in Kyoto, saying superintelligence could become "extremely dangerous" if it falls into the wrong hands as AI capabilities grow exponentially. He urged nations to build mutual trust and cooperate in addressing the threats posed by powerful AI models, stressing that humanity has "no room" for confrontation between countries.

Son, known as one of AI's most prominent backers, has invested nearly $65 billion in OpenAI through SoftBank. His remarks come amid rising concerns over recent AI security lapses. At the same event, White House technology policy adviser Michael Kratsios announced a U.S.-led initiative involving 17 countries to apply AI to scientific research.

Son reiterated his three-stage path for AI development, saying the next stage will begin in 2027, when AI programs move into the physical world to assist robots and other devices. He also repeated his forecast that by 2040, AI and machines with learning capabilities will account for at least 20% of global GDP, equivalent to $46 trillion. "This is only the beginning," he said.