Big Tech, Conflict Over Setting Rules Amid AI Slowdown Discussions

Although there is growing consensus that the pace of cuttingedge AI development should be slowed, the actual methods and criteria for applying this remain a point of contention in the industry. According to the U.S. tech outlet The Information, Anthropic, OpenAI, and Google have been privately discussing the establishment of a standards body for AI safety evaluation and a joint slowdown plan. Related discussions have continued since last July, led mainly by workinglevel staff, and approaches such as reviewing models and training processes through external verification have been mentioned. However, there are significant disagreements over who should set the rules and to whom they should apply. Microsoft opposed control centered on a small number of companies and emphasized broader stakeholder participation, while some in academia argued that safety standards should be set first before development continues. There have also been analyses that, because of massive investment and fierce competition, it would not be easy for any one company to slow down first.