News Summary

OpenAI has announced potential adjustments to its AI safety requirements if competing AI labs release high-risk systems without adequate safeguards. The update to OpenAI’s Preparedness Framework emphasizes maintaining a high level of protection even if rivals opt for less stringent measures. OpenAI’s decision comes amidst increasing competitive pressures in the AI industry, where rapid deployment of models is often prioritized. The company has faced criticism for allegedly lowering safety standards in favor of speedier releases, a claim OpenAI disputes. The updated framework also outlines a focus on automated evaluations to accelerate product development while maintaining safety protocols. Additionally, OpenAI now categorizes AI models based on their potential risk, focusing on ‘high’ capability and ‘critical’ capability models, with each requiring specific safeguards. These changes mark the first updates to the Preparedness Framework since 2023.