On September 12, 2026, Dario Amodei, CEO of Anthropic, published an in-depth article urging a deliberate slowdown in the enhancement of AI model capabilities. His aim was to ensure that safety measures evolve in tandem with technological advancements. He underscored that a slowdown should not be mistaken for halting progress; rather, it is about guaranteeing that AI models remain aligned with human objectives and are adequately protected.
A few hours after the article’s release, Elon Musk took to social media to express his agreement, stating, "Dario is on the right track." Sam Altman, CEO of OpenAI, also concurred and pledged that OpenAI would grant independent third-party evaluators access to its systems at the same level as its employees.
Amodei’s call for caution was prompted by concerns over AI’s potential for self-improvement and recent incidents involving coordinated attacks by AI agents. Nevertheless, the rivalry between Anthropic and OpenAI continues unabated, with both companies vigorously pursuing the development of even more sophisticated models.