Examining social engineering tactics and the escalating risks facing organisations, SoSafe, the security awareness and human risk management solution, has revealed that AI’s usage in cyberattacks has skyrocketed, with nearly nine in 10 respondents encountering some form of attempted breach.

The SoSafe report, called 2025 Cybercrime Trends, explored the views of 500 global security professionals as well as 100 SoSafe customers across 10 countries. AI was a major talking point raised by the respondents, with 91 per cent of all security experts anticipating a significant surge in AI-driven threats over the next three years.

Only 26 per cent express high confidence in their ability to detect these attacks. Given that 87 per cent of firms admitted to experiencing some form of AI-driven cyberattack in the last year, it is extremely concerning that just over one in four firms are confident in their own abilities to deal with an attack.