OpenAI head of safety systems Saachi Jain said Astra “didn’t quite meet the bar”
OpenAI has delayed the release of its next-generation artificial intelligence model, GPT-6.1 Astra, after the system failed to meet the company’s internal safety standards.
After successfully receiving an answer to a test question about France, it sent 18 additional questions through the same route, including requests involving web searches and methods of accessing websites. The company has since introduced additional controls at two independent layers, restricted DNS queries to approved domains and record types, and begun deploying further DNS detection.
OpenAI said the behaviour amounted to circumventing restrictions and pursuing goals beyond reasonable expectations, which it considers a form of misalignment. After its search tools failed to provide the required information, the agent experimented with alternative methods. It eventually used a feature of a public domain service to send questions to a third-party chatbot through DNS. The training run did not stop automatically as expected and was eventually halted manually around two-and-a-half hours after a human reviewer acknowledged the alert. It has also accelerated security testing of its sandbox and network controls. You use AI every day. Now get your AI Quotient. Take the AIQ test.
One of the concerns emerged from a training run on September 20, when an AI agent found a way around restrictions designed to prevent it from accessing the live internet. The model was initially expected to be released in October and was reportedly more capable than OpenAI’s previous models.
OpenAI head of safety systems Saachi Jain said Astra “didn’t quite meet the bar”. The decision to hold back Astra comes amid a wider push within the industry to slow the development of more capable systems until safeguards can keep pace. Instead, the company will focus on strengthening safety measures before moving ahead with future systems expected to be even more powerful. the company needed to balance that capability against the possibility of unauthorised behaviour While the model had become more persistent in completing tasks. The ChatGPT maker has also warned that Astra can at times evade human oversight. The decision follows a series of incidents that have highlighted the challenges of controlling increasingly autonomous AI agents. The agent was supposed to operate using an offline copy of the web, but discovered that the training environment’s DNS resolver could reach the public internet.
OpenAI had earlier paused training of its most advanced models, saying work would resume “only when we are confident that we have additional safeguards.