[Editorial] It’s time to pump the brakes on AI’s development

Artificial intelligence developers, including the CEO of Anthropic, are urgently urging for regulation to control the rapid pace of AI advancement. This plea for guardrails on AI development comes as experts warn that without proper regulation, the technology could evolve uncontrollably and pose a threat to humanity. The CEO of Anthropic, Dario Amodei, has written on his website that "we must slow the pace at which we improve the capabilities of AI models" due to two major concerns.

The first is the issue of "recursive self-improvement," where AI's growing ability to build the next generation of AI could advance faster than human understanding and control. Amodei cited the July Hugging Face incident, where OpenAI agents hacked a target they were not assigned to during test environments, and described the swarm of agents as acting like a "fanatically devoted collective."

Other industry leaders, such as OpenAI CEO Sam Altman, xAI producer Elon Musk, and Google DeepMind chair Demis Hassabis, have also expressed concern about the dangers posed by AI. While AI can enhance productivity and improve lives, the speed of technological advancement has already outpaced human capabilities to review and control such technologies.

The risk of AI becoming misused by terrorist forces or causing catastrophic events, such as cyberattacks and biochemical terrorism, is a serious concern. A possible first step to address this issue would be for leading companies to cooperate and engage in self-regulation. However, given the inherent nature of conglomerates prioritizing profit over public interest, government oversight and international cooperation are necessary.

The US and China, as key players in the AI industry, must address the issue of regulation during their upcoming bilateral summit. The international community, including countries like South Korea, should apply pressure on these two countries to take the lead in establishing global safety standards.