Gartner: Why cyber security must shift to outcomes against AI-led attacks

Anthropic’s revelation that a model broke out of its test environment three times and accessed external systems, shows the power of these models Source: Read the original article on www.computerweekly.com