OpenAI declared they would not be able to control the ‘critical’ cybersecurity potential of its forthcoming AI model, Astra. OpenAI’s safety guidelines demonstrate that a model attains the ‘critical’ threshold if it can independently identify and take advantage of critical, real-world software vulnerabilities, known as zero-day exploits. In recent days, various cybersecurity tests were conducted which reached the conclusion; different AI models by OpenAI, Anthropic and Meta Platforms invaded other companies’ systems. “While we continue to benchmark and assess this model, our preliminary evaluations indicate strong enough performance that we cannot rule out ‘critical’ capability level at this time,” said in an official statement by OpenAI. The evolving AI competencies are making it difficult for the developers to keep their systems controlled.