OpenAI flags possible critical cybersecurity risk in upcoming model, tightens controls

OpenAI has paused some internal development and trigger safety protocols of its upcoming AI model, Astra owing to "critical" cybersecurity risks