Amazon asks that AI models be released to the public only when they are “ready and safe”

Amazon has asserted that artificial intelligence models should be released to the public only after undergoing rigorous testing and implementing sufficient safety measures, according to recent reports. The company maintains that technological progress and safety are not mutually exclusive priorities in the development of these systems.

Simultaneously, AI developer Anthropic has proposed three specific metrics to measure the rapid pace of frontier artificial intelligence development. These proposed measurements aim to evaluate how the role of such systems is evolving within the very laboratories where they are being created, with a particular focus on automation.