Cyber Incidents hit record high in 2025 as AI weaponisation drives 27% jump

Phishing remains a dominant threat, accounting for 57% cases.

Hong Kong’s cyber incidents reached a record high of 15,877—a 27% year-on-year increase—in 2025, driven largely by the weaponisation of artificial intelligence (AI).

Findings from the Hong Kong Computer Emergency Response Team Coordination Centre (HKCERT) revealed that phishing attacks remained the most prominent threat, accounting for nearly 57% of total cases. Incidents involving vulnerable systems also rose to 2,328 cases, representing a more than 3.5-fold increase, whilst botnet-related incidents accounted for 18%.

Attack delivery methods have expanded beyond traditional email to include social media and instant messaging platforms, which accounted for 34% of reported incidents, as well as cryptocurrency platforms at 18%.

Key risks include AI-driven and autonomous attacks, data exposure arising from weak AI governance policies, and security gaps within third-party supply chains.

The report also highlights the dangers of cloud-based single points of failure and the growing threat posed by connected AI devices.

It further revealed a divide in preparedness amongst enterprises, with only 26% of small and medium-sized enterprises (SMEs) having dedicated cybersecurity personnel, compared to 59% of large enterprises. Whilst 48% of SMEs have adopted email security measures, this remains well below the 79% adoption rate amongst large enterprises, the report said.

For privileged access management, only 29% of SMEs have implemented such controls, compared to 60% of large enterprises.

Similarly, the adoption of advanced remote security remains limited, with just 31% of SMEs implementing these measures, compared to 67% of larger organisations—underscoring the need for greater support for SMEs in driving technological upgrades.

To mitigate these emerging risks, HKCERT recommended that organisations prioritise the appointment of dedicated cybersecurity monitors and implement formal AI governance frameworks.

It also stressed the importance of fostering a cross-functional security culture to combat phishing, delivering ongoing staff training to minimise human error, and strengthening technical infrastructure through multi-factor authentication, robust encryption, and proactive intrusion detection.