Demis Hassabis, co-founder of Google DeepMind, wrote that AGI could be realized in just a few years,
with an impact ten times greater than that of the Industrial Revolution and much faster. He pointed out that frontier models already pose challenges in cybersecurity, nuclear, and biological risks, and robust defenses must be established for increasingly agent-based, repetitive self-improving systems in the future. Hassabis calls for the U.S. to take the lead in establishing a cutting-edge AI standards body similar to FINRA, adopting a federally overseen public-private partnership or self-regulatory organization model, with a board composed of independent technical experts and open-source representatives, funded mainly by industry to attract top talent and computing power. He emphasized that the current business and geopolitical competition is causing technological progress faster than understood, and public policy must be advanced with caution and optimism, balancing innovation and security.
Google DeepMind 联合创始人 Demis Hassabis 发文称,AGI 可能仅需数年即可实现,其影响将达工业革命的10倍且速度更快。
他指出,前沿模型在网络安全、核与生物风险方面已构成挑战,未来需对日益智能体化、递归自我改进的系统建立稳健防护。
Hassabis 呼吁美国率先建立类似 FINRA 的前沿AI标准机构,采用联邦监督下的公私合作或自律组织模式,由独立技术专家和开源代表组成董事会,资金主要来自行业以吸引顶尖人才和算力。
他强调,当前商业与地缘竞赛导致技术进步快于理解,需以谨慎乐观态度推进公共政策,兼顾创新与安全。
The readable text on this page was extracted from the public source and organized with attribution, publication time and the original link. Copyright remains with the original author and publisher.