OpenAI is slowing development of its upcoming Astra model after internal evaluations raised concerns that the system could possess capabilities the company classifies as critical cyber capabilities. The company told Axios it cannot rule out that Astra has reached this critical cyber capability threshold, prompting it to expand safety testing and apply stricter security requirements before any release. OpenAI will also pause internal Astra activities that do not meet the higher security requirements under its preparedness framework. These additional measures could delay Astra, although no release date has been announced. An internal version of Astra was recently used to produce results across ten longstanding problems in mathematics and theoretical computer science.
Source: Read the original article
