AI safety push creates internal rift at OpenAI, Anthropic

OpenAI: OpenAI develops advanced AI models and tools for general use. In the context of the news, its leadership's emphasis on additional safety protocols amid model development has contributed to internal divisions over priorities and timelines. Recent industry discussions highlight the company's involvement in collaborative safety talks with peers. Anthropic: Anthropic builds AI systems with a focus on safety and alignment research. The news describes how its executives' safety initiatives have generated internal tensions similar to those at OpenAI, including debates around development pace. The company has publicly advocated for measures like independent evaluators in recent weeks. Employee Concerns: Recent researcher departures at frontier AI labs have amplified calls for greater caution in advancing capable systems due to risks like recursive self-improvement. Leadership Alignment: CEOs from leading AI companies have publicly backed proposals for third-party model evaluations and potential pacing of frontier development. Safety Collaboration: OpenAI, Anthropic, and Google have been engaged in multi-week discussions on establishing industry-led AI safety standards and practices.