Seoul: Warning voices from within the U.S. artificial intelligence (AI) industry, from researchers to executives, have raised the alarm about safety challenges related to the technology. For Korea, where President Lee Jae Myung has repeatedly pushed for the country to become one of the world's top three AI powers as a goal for his administration, these warnings signal an urgent need for shared safety regulations.
According to Yonhap News Agency, over the weekend, Anthropic CEO Dario Amodei, who has been warning about AI safety risks, called on companies to slow the pace at which they are developing the technology. He has repeatedly expressed concern over the dramatic advancement of AI and the growing possibility of recursive self-improvement, in which AI systems are able to autonomously develop a new generation of AI. Amodei stated, "I believe that if slowing down bought us even an extra year or two before models reach critical levels of capability, and we used that time to advance alignment, we could greatly reduce the risk that something goes seriously wrong." Support from rivals like Elon Musk, who owns SpaceXAI, and OpenAI CEO Sam Altman has given Amodei's warning even more weight.
Former OpenAI and Anthropic researcher Jacob Coxon, who resigned over concerns about AI safety, warned that firms "are racing straight to self-improving superintelligence and gambling with our lives." In a post on X (formerly Twitter), Coxon said those building AI earnestly "believe that it could kill us all by the end of the decade."
The call is for a coordinated slowdown among AI firms and governments, not stopping research altogether. Some have even voiced skepticism about the motivations behind Amodei's warnings, suggesting AI firms intend to use the slowdown to shore up their existing leads in the sector.
The warnings also come amid competition between the world's leading powers. U.S. President Donald Trump recently dismissed concerns over the safety risks of AI, attributing them to "very negative forces." He asserted, "We're leading China in AI. We're the most sophisticated country in the world, and frankly I want to keep it that way because whoever wins AI wins."
China has also reacted to Amodei's remarks, with Chinese foreign ministry spokesperson Guo Jiakun stating, "Fearmongering, confrontation and vicious competition will only disrupt the process of global AI governance which serves no one's interest."
These comments illustrate the political tension caused by the AI race between the U.S. and China, a geopolitical factor further complicating responsible AI development. The hope is that Trump and Chinese President Xi Jinping will tackle this issue during their meeting on Sept. 24. Korea should keenly observe the outcome.
While AI development has shown numerous benefits, continuous concerns linger about job displacement and the potential for AI to go rogue. In July, an incident involving autonomous AI agents from OpenAI breaching the systems of Hugging Face, an AI infrastructure company, highlighted these risks, though the industry assessed that no lasting damage was done.
Anthropic has proposed that independent evaluators should have access to AI labs to assess risk, firms should set common safety standards, and democratic countries should pursue international coordination with China and other governments.
Korea should reassess its current AI governance system to ensure it has sufficient safeguards in place as it develops its sovereign AI model. As a democratic export powerhouse with strategic ties to China and a longstanding alliance with the U.S., Korea needs to ensure it is not caught in the cutthroat AI competition between the two countries. Efforts should be made to facilitate coordination between the U.S. and China toward responsible AI investment.