AI Can Now Execute 73% of Expert-Level Cyberattacks. Here’s What IT Leaders Need to Do Now.
On April 7, 2026, Anthropic announced a model so capable they refused to release it publicly. Claude Mythos, their most advanced frontier AI, was deemed too dangerous for open access because of one thing: it can hack. Not in the vague, theoretical sense that security researchers worry about. In the “completed 73% of expert-level capture-the-flag… Continue reading AI Can Now Execute 73% of Expert-Level Cyberattacks. Here’s What IT Leaders Need to Do Now.
