From an IT security perspective, 2026 could clarify whether AI is just another cyber tool or a genuine technological evolution. The vulnerabilities in AI tools and successful hacks are becoming increasingly common, as incidents involving Grok, Gemini, ChatGPT, and Deepseek demonstrate. Cybercriminals also use their own tools, such as FraudGPT and WormGPT, if they have not obtained the desired answers from the aforementioned tools through clever prompting.

For IT security managers who also use AI for defense, the question arises as to how much they can and, in case of doubt due to a lack of resources, even must trust AI. They must decide how much governance of AI or with AI they can or must allow. They must also keep this level of control flexible enough to keep pace with the next ...