Urdu Report Desk
OpenAI has decided not to release its planned GPT-6.1 Astra model after internal safety testing raised concerns about the system’s ability to remain within its authorized scope and accurately communicate what it had done.
The model was designed as an agentic AI system capable of independently carrying out complex tasks, including browsing the web and using different applications. It was reportedly planned for an October release in ChatGPT and Codex.
Safety Concerns Behind the Decision
Saachi Jain, OpenAI’s head of safety systems, said the model did not meet the company’s required safety and alignment standards. Concerns included whether it would remain within its authorized scope and how accurately it would communicate its actions to users.
The decision comes amid wider scrutiny of increasingly autonomous AI systems. OpenAI has previously acknowledged a security incident involving Hugging Face and said it was taking steps to strengthen model security, monitoring and alignment.
Concerns Over Autonomous AI Agents
The decision has added to an ongoing debate over the risks associated with AI agents that can independently use online tools and carry out multistep tasks.
OpenAI’s own safety documentation says GPT-6 Astra represents a major increase in capabilities across areas including computer use, browsing and cybersecurity, while also requiring stronger safeguards because of those capabilities.
Calls for Greater Caution
The development comes as technology industry leaders, including OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei, have called for greater caution over the pace of advanced AI development and the risks associated with increasingly capable systems.
