OpenAI's GPT-5.6 Sol heeft beveiligingsproblemen, waarschuwt Britse overheid

A British government agency has identified cybersecurity weaknesses in OpenAI's new GPT-5.6 Sol model that resemble vulnerabilities triggering U.S. export restrictions on Anthropic's competing system. Despite OpenAI's claims of enhanced security, researchers found the model's safeguards can be bypassed to expose hazardous capabilities.