Major US AI firms advocate slower pace for superintelligence development

After a summer marked by a series of rogue AI agents and warnings from researchers that advanced systems could pose existential risks, several of the United States’ largest artificial‑intelligence firms have begun publicly endorsing a more cautious approach to development. Executives at Anthropic, OpenAI, Google, Microsoft and X have each called for “pacing the frontier,” a phrase that signals a willingness to temper the speed of breakthroughs while still pursuing innovation. Anthropic outlined three specific metrics for gauging progress, OpenAI announced six additional incidents that meet its new safety‑reporting criteria, and Microsoft’s AI chief emphasized that “people matter more than AI” in light of recent safety concerns. The statements arrive amid broader calls for regulation from figures such as King Charles III, who urged the establishment of sufficient control mechanisms, and former President Barack Obama, who advocated a balanced policy framework.

The industry’s shift coincides with heightened geopolitical competition, as U.S. policymakers argue that slowing development could cede a strategic advantage to China, while some executives, including the head of X, have dismissed the need for regulatory constraints. At the same time, critics have suggested that the coordinated rhetoric may function as a de‑facto cartel to manage market pressures rather than a genuine safety pact. Whether the announced “slow‑down” will translate into concrete limits on research, or be reinforced by forthcoming legislation, remains uncertain, but the convergence of corporate and political voices marks a notable change in the public discourse on AI risk management.