OpenAI to Submit ChatGPT Models for US Gov Review

OpenAI has confirmed it will comply with a US government executive order regarding AI model assessment. This order asks AI companies to allow the US government to assess their models' capabilities before public release. The order, signed on Tuesday, focuses on early access to advanced AI systems. It requests companies provide access to their latest AI models 30 days before their public launch. OpenAI plans to participate in this review framework, submitting its next-generation AI models for a structured benchmarking process. This process will assess advanced cyber capabilities and help determine if a system should be classified as a "covered frontier model" for additional oversight. Earlier versions of this policy were more stringent, but were revised after public criticism and industry concerns. This move highlights ongoing efforts to monitor AI safety and security risks globally.