Now Reading: OpenAI pauses Astra AI over critical cyberattack fears
-
01
OpenAI pauses Astra AI over critical cyberattack fears
Times Now reports that OpenAI has paused some internal activities around its upcoming AI model, Astra, after internal tests suggested it may have reached a “Critical” level of cybersecurity capability. Astra is…
The Story in Brief
Times Now reports that OpenAI has paused some internal activities around its upcoming AI model, Astra, after internal tests suggested it may have reached a “Critical” level of cybersecurity capability. Astra is an agentic AI model that can perform tasks, use tools, and achieve goals with minimal human intervention. In testing, the model showed “significant advancements in agentic coding and cybersecurity.”
OpenAI’s Preparedness Framework defines a “Critical” cybersecurity capability as an AI that could discover and exploit serious weaknesses in crucial computer systems with little or no human help. The company warned that cybersecurity is changing rapidly, and models can both strengthen defences and enable attacks at unprecedented speed. While OpenAI has not confirmed that Astra has crossed the threshold, it is strengthening security measures before allowing some internal activities to resume.
The Indian Opinion
The alarm around Astra AI risks feeding a familiar cycle: tech firms hype their own breakthroughs, then pause them to show responsibility, while the public is left wondering if real danger or marketing is driving the stop. OpenAI has not confirmed Astra crossed the critical threshold, it only says it cannot rule it out. That is a careful hedge. The real test will come when the company publishes its full evaluation. Until then, treat claims of “unprecedented” cyber capabilities with the same scepticism as the model’s own coding tricks.
Source: timesnownews.com
This story was synthesised by AI from the source linked above.
