Newsletters
Serious cyber incidents worldwide reached a record high in the first half of 2026, up 71% from a year earlier, and researchers warn that artificial intelligence could soon make attacks bigger and faster, according to a new report from Italy’s national cybersecurity association, Clusit.
The group counted 4,911 serious incidents globally between January and June, an average of 819 a month.
This is 105% more than in the previous six months and already equal to 93% of all incidents recorded in the whole of 2025. The updated report was presented in the Italian city of Verona.
The researchers pointed to early signs of what they call "agentic hacking," in which AI programs carry out attack tasks with growing independence.
They warned that coordinated groups of AI agents could eventually launch fully automated attacks, changing tactics on the fly and combining methods to exploit weaknesses in company systems.
To counter this, they called for clearer lines of responsibility, flexible governance and defenses able to respond quickly and automatically to AI-driven threats.
Italy recorded fewer known incidents, 245 against 280 in the first half of 2025, making up about 5% of the global total. But the share rated "high" severity jumped by nearly 20 percentage points to 57%.
Manufacturing was the hardest-hit sector as financially motivated attacks on businesses rose.
Hacktivism, meaning hacking meant to advance political, social, or ideological causes, declined. Malware was behind 43% of Italian incidents, up from 22.7% last year.
Researchers said stronger defenses introduced under Italian cybersecurity law and the European Union’s NIS2 directive could partly explain the drop in hacktivism.
"The data tells us that the nature of risk is changing," Clusit President Anna Vaccarelli said. She urged organizations to treat cybersecurity as essential to keeping their business running and as a shared responsibility that supports national resilience.
Set as preferred source
Set as preferred source
1 min read
1 min read
