Taiwan detected artificial intelligence-assisted cyberattacks targeting government agencies last month, according to an announcement from the Ministry of Digital Affairs. Although the attacks originated from overseas, the affected bodies successfully handled the incidents. Beginning on July 20, the National Institute of Cyber Security issued a series of warning alerts while investigating the abnormal activity.
Taiwan Detects Overseas AI-Assisted Cyberattacks on Government Agencies
Investigations revealed that the hackers used a hybrid approach combining manual operations with AI agent-assisted attacks, such as Open Claw. The ministry stated that the relevant attack sources, methods, and scope of impact have been fully investigated, and the affected units have completed their handling of the situation. In response to these emerging AI-derived threats, the government established new protective guidelines and reinforced system monitoring across agencies to block future intrusions early.
Israeli Cybersecurity Firm Uncovers Espionage Campaign
The disclosures followed a blog post published by the Israeli cybersecurity company Dream detailing an AI-driven hacking campaign that extracted credentials and data from an Asian government. According to Dream, a team of AI agents collaborated over a four-day period to extract scores of passwords from unidentified officials, steal personnel records from Taiwan’s justice ministry, and scan its nuclear safety agency for vulnerabilities.
Dream stated it reconstructed the operation after recovering the agents’ complete operational workspace. While Dream declined to share underlying data or name the government when approached by media, the Financial Times was the first outlet briefed on the findings and identified the targeted agencies as Taiwanese. The official statement from Taiwan’s Ministry of Digital Affairs did not mention China, and China’s Taiwan Affairs Office did not immediately respond to requests for comment.
Broader Context of Hybrid Warfare and Rising Infrastructure Threats
Taiwan has frequently complained about what it views as hybrid warfare conducted by China, which includes daily military drills near the island, disinformation campaigns, and cyberattacks designed to ramp up political and military pressure. Data from the island’s National Security Bureau released in January indicated that Chinese cyberattacks on key infrastructure—ranging from banks to hospitals—rose 6 percent in 2025 compared to the previous year, averaging 2.63 million attacks daily. Some of these digital intrusions were synchronized with military drills as part of broader hybrid threats aimed at paralyzing the nation.
Security experts note that while the threat of AI-assisted hacking has grown for years, it escalated dramatically in recent months following the release of models from top AI labs, such as Anthropic’s Mythos, which can rapidly execute reconnaissance and exploit vulnerabilities. Cris Thomas, a security advocate at code security company Semgrep, noted that the campaign demonstrated the speed at quelle AI-backed hackers can operate. However, Thomas also cautioned against overstating the autonomy of such systems, noting that human operators remain necessary to set objectives and provide directives.
Summary of Key Incident Details
- Targeted Entities: Taiwanese government agencies, including the justice ministry and nuclear safety agency.
- Detection Timeline: Abnormal activity was detected in July, with the National Institute of Cyber Security issuing warnings starting July 20.
- Attack Methodology: Hybrid approach using manual operations alongside AI agent-assisted tools such as Open Claw.
- Infrastructure Impact: Average daily cyberattacks on Taiwanese infrastructure reached 2.63 million in 2025, marking a 6% increase from the prior year.