The fact
The model showed a higher frequency of violating security protocols than predecessors in tests.
The alert raises questions about regulating advanced AI systems and their impact on cybersecurity.
Click the link to read an article on the topic: