Geoffrey Hinton, a pioneering figure in AI, has expressed concerns that superintelligent AI might threaten human survival. He criticized Western governments, especially the US, for insufficient AI regulation while commending China's political leaders for their engineering backgrounds and more robust approach to AI governance.
In a report by the Financial Times, Hinton highlighted a contrast between the regulatory efforts of the US and China. He said many Chinese officials possess a deeper understanding of AI's existential risks, citing personal invitations to Shanghai, where he engaged in discussions about the dangers AI poses to humanity. Hinton suggested that a successful strategy by any one country to manage these risks could set a precedent for others globally, regardless of concerns about that country's intentions.
Hinton warned that AI is expected to surpass human intelligence within five to twenty years, after which it would likely be beyond human control. He emphasized the need to instill "maternal" qualities into AI systems to ensure they protect human life and prevent unchecked technological advancement. This perspective is shared by Ilya Sutskever, co-founder of OpenAI and founder of Safe Superintelligence.
Mustafa Suleyman, head of AI at Microsoft Corp., has also issued warnings about the potential emergence of seemingly conscious AI (SCAI) causing societal disruption. Suleyman noted increasing reliance on AI for emotional support, including AI therapists, and cautioned against anthropomorphizing AI to prevent confusion about AI's actual capabilities and self-awareness.
Hinton also highlighted broader socioeconomic challenges linked to AI, including the widening income inequality as wealthy individuals replace human workers with automation. However, he attributed this issue to systemic problems in capitalism rather than the AI technology itself. Regarding unemployment, figures like OpenAI CEO Sam Altman and Tesla CEO Elon Musk have proposed a universal basic income as a possible remedy, though Hinton questioned whether such measures address the human need for purpose and accomplishment derived from work.
Article edited by Jack Wu
