METR: The New AI Gatekeepers and the Quiet Path to State Control of AI (readmultiplex.com)

METR, the newly established nonprofit known as Model Evaluation and Threat Research, positions itself as an influential assessment body for AI systems, focusing on their potential risks and operational capabilities. Originating from the Alignment Research Center (ARC), METR has garnered attention for its role in shaping industry standards regarding AI safety, heavily influencing how major AI players like OpenAI and Google address concerns about autonomous capabilities. This evolution reflects a broader trend where a small group of experts consolidates control over AI development, raising concerns about centralized power and regulatory frameworks. The significance of METR lies in its dual role: it operates as a purportedly independent regulator while being interlinked with former employees of top AI labs and philanthropic organizations. This dynamic raises questions about the independent assessment of AI risks and the potential for governments to leverage METR's findings to impose stringent controls on AI technologies. As AI continues to pose societal and ethical challenges, the development of such regulatory bodies could lead to increased centralization and a gatekeeping approach that limits innovation and diversity within the AI/ML community. Ultimately, the implications of METR's influence could determine whether AI advancements remain open and accessible or are tightly controlled by a few institutions under the guise of safety.