AI-driven cyberattacks surge in Taiwan; semiconductor sector targeted
Taiwanese organizations faced a surge in AI-driven cyberattacks from late 2025 to mid-2026, with semiconductor and government sectors hit hardest, averaging nearly 4,000 attacks per week. Criminal groups have adopted AI as a commercialized attack tool, automating fraud and jailbreak techniques while drastically shortening the time from vulnerability discovery to exploitation.
What happened
From November 2025 to May 2026, Taiwanese organizations faced an average of 3,974 cyberattacks per week, 92% higher than the global average, according to Check Point Research's threat intelligence report.
The semiconductor, government, and industrial manufacturing sectors were the primary attack targets. High-risk prompts increased 97% year-on-year, indicating AI has become a new security vulnerability as attackers integrate it into attack workflows and exploit processes.
Reported impact
The time from vulnerability discovery to exploitation by attackers has compressed to hours in the AI era. Ransomware victims in Taiwan grew from 8 cases in the previous quarter to 26 cases in the first quarter of 2026.
Index analysis
AI-driven attacks have shifted from nation-state tools to standardized, repeatable processes commercialized for criminal use, forming 'AI Attack Factories.' Agent profiles are being weaponized as persistent jailbreak attack channels, and AI fraud workflows including model selection, jailbreak operations, and content generation have become highly automated, lowering attack barriers.
Governance controls
Check Point recommended that enterprises inventory all AI applications, establish objective risk rating systems, build AI risk inventories, establish governance committees, and strengthen shadow AI management and employee training to address uncontrolled 'shadow AI' within organizations.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.