Cybercriminals are not (yet) the AI supervillains one might fear

New analyses of cybercrime paint a more sober picture than headlines about AI hackers often suggest.

Financially motivated criminals have access to advanced AI tools, but very few use them. It's not the ability that's lacking, but the organizational framework: it has to pay off before it makes sense to invest time in new methods.

In practice, this means that most attacks are still fairly mundane in their execution. Many cybercriminals buy ready-made vulnerabilities (so-called zero-days) instead of developing advanced AI attacks themselves, simply because it's cheaper and faster.

For you and your company, this isn't a free pass to relax. But it's also no reason for acute panic over AI-driven super threats. Ordinary IT security (updated systems, awareness and good routines) still covers the vast majority of attacks. Keep an eye on developments: if concrete examples of widespread AI use in actual attacks start to appear, it's time to reassess your preparedness.