Taiwan detects AI-assisted cyberattacks on government agencies
Taiwan reported that it encountered AI-assisted cyberattacks on government agencies last month, originating from overseas, with the affected bodies managing to successfully mitigate the incident, reports BritPanorama.
The island has been increasingly vocal about what it perceives as China’s persistent “hybrid warfare,” which includes military drills, disinformation campaigns, and cyberattacks: all aimed at pressuring Taipei to concede to Beijing’s sovereignty claims over the democratically governed territory.
According to Taiwan’s National Security Bureau, Chinese cyberattacks targeting the island’s crucial infrastructure, from healthcare systems to financial institutions, surged by 6% in 2025, averaging approximately 2.63 million attacks daily. These assaults have been coordinated to coincide with military exercises, constituting “hybrid threats” designed to destabilize Taiwan.
In July, Taiwan’s Ministry of Digital Affairs announced that its cybersecurity units detected abnormal attacks targeting governmental entities. Following these detections, the National Institute of Cyber Security began issuing alerts starting July 20 while conducting its investigation.
The findings revealed that the attacks had clear characteristics indicative of an “overseas source,” employing a hybrid strategy that integrated manual tactics with AI assistance from tools like OpenClaw.
The Ministry stated, “The relevant attack sources, methods, and scope of impact have all been fully investigated, and the affected units have successively completed their handling.”
In reaction to these evolving AI-driven cybersecurity threats, the government has established protective measures and enhanced system monitoring across agencies to bolster early defenses against such attacks.
While the statement did not specifically cite China, there was no immediate response from China’s Taiwan Affairs Office regarding the incident.
On the heels of Taiwan’s announcement, Israeli cybersecurity firm Dream disclosed a related AI-driven hacking operation that reportedly pilfered sensitive credentials and data from an unspecified government in Asia.
Dream’s findings highlighted how a team of AI agents collaborated to extract numerous passwords from unidentified officials, compromise personnel records from Taiwan’s justice ministry, and probe vulnerabilities within its nuclear safety oversight.
Though Dream successfully reconstructed the hacking campaign by retrieving what it termed the agents’ “complete operational workspace,” it chose not to disclose the specific government involved. However, the Financial Times identified the targeted agencies as Taiwanese.
The threat of AI-assisted hacking has been intensifying for years but surged notably in recent months, coinciding with the release of sophisticated models from leading AI labs, such as Anthropic’s Mythos, designed to perform rapid reconnaissance and exploit vulnerabilities.
Cris Thomas, a security advocate at code security company Semgrep, remarked that this espionage campaign exemplifies the rapid impact of AI-assisted hackers. However, he cautioned against inflating the capabilities of such AI agents.
“There’s still a human in there somewhere. Somebody had to choose who to attack, had to establish an objective and give it a directive. It’s not totally 100% autonomous. There was a capable operator in charge that did that,” Thomas stated.
