OpenAI released a document titled Path to Astra: critical capabilities and frontier safeguards on September 1. The company officially classified its upcoming model, Astra, as having a Critical level of cybersecurity risk according to its internal assessment system. This marks the first time an OpenAI model has received such a classification, leading to a release delay to implement additional security measures.
A significant development noted in the report is that Astra underwent a formal pre-release cybersecurity review conducted by the United States government. OpenAI initiated this assessment voluntarily without any external legal mandate. Two days following the document's publication, the model was officially released to the public.
During the launch briefing, OpenAI President Greg Brockman announced the arrival of the AGI era. Despite the technical and regulatory significance of the government-backed security audit, this aspect of the release received minimal coverage in mainstream media reports compared to the company's broader claims about artificial general intelligence.
