Kaspersky predicts that Artificial Intelligence (AI) will intensify both defense capabilities and threats in the field of cybersecurity in 2026, with direct implications for companies and public entities.
The evolution of language models and content generation tools is transforming the way attacks are designed and executed, requiring greater investment in prevention, training and technology.
The consultant identifies deepfakes increasingly realistic — especially with advances in audio — and a lower barrier to entry, which makes it easier for cybercriminals to use them. A situation that should lead to an increase in operational costs for companies, which will have to intensify employee training and adopt stricter internal policies to mitigate risks of fraud and social engineering, according to Kaspersky.
The company adds that the models open‑sourcewhich increasingly come closer to proprietary solutions, reduce controls over their use and widen the risk window, pressuring organizations to review data governance strategies and compliance. In parallel, the increasing difficulty in distinguishing legitimate from synthetic content promises to complicate automatic detection and increase the burden on technological solutions and monitoring services.
Kaspersky also predicts that AI will support various phases of attacks — from code engineering to evidence hiding — while becoming a central tool for security teams, automating analyzes and reducing manual tasks. This implies a change in investment: less time on routine operations, but more spending on advanced tools and specialized training, according to the consultant.