Taiwan Detects Unusual AI-Assisted Cyberattacks Targeting Government Agencies

TAIPEI:

Taiwan has detected a series of unusual cyberattacks targeting government agencies that allegedly used artificial intelligence to automate parts of the hacking process, according to the island's Ministry of Digital Affairs.

The attacks, which originated from overseas and were first detected on July 20, targeted government systems and later expanded to include Taiwan's nuclear safety authority and at least seven energy companies, the ministry said.

Israeli artificial intelligence security company Dream told the Financial Times that the attackers used open-source AI agents to develop an automated hacking tool capable of operating like a coordinated cyber team.

Dream described the incident as unprecedented in its experience.

85 Government Accounts Compromised

According to Dream, the attackers gained access to at least 85 government user accounts and collected information relating to more than 2,500 employees.

Taiwan's cybersecurity monitoring authorities detected the activity, while the National Institute of Cyber Security issued several warnings during the investigation, according to the Ministry of Digital Affairs.

The ministry said the investigation into the attack's origin, methods and scope had been completed and that affected agencies had taken necessary measures.

Taiwanese authorities have also developed defensive guidelines to address emerging AI-enabled cyber threats and have strengthened cybersecurity monitoring systems to improve early detection and prevention.

Suspected Links to China

Taiwanese authorities have not publicly attributed the attacks to China.

However, the Financial Times, citing Dream, reported that hackers linked to China were suspected of being behind the operation.

Dream did not identify a specific hacking group. The company said, however, that internal communications associated with the operation contained Simplified Chinese, which it considered a possible indication of links to China.

Taiwan has for years accused China of conducting what it describes as a form of hybrid warfare, including military exercises around the island, disinformation campaigns and cyberattacks.

According to Taiwan's National Security Bureau, Chinese cyberattacks targeting critical infrastructure increased by 6% in 2025 compared with the previous year. The bureau said critical facilities, including hospitals and banks, faced an average of about 2.63 million cyberattack attempts a day, with some attacks reportedly coordinated with Chinese military exercises.

China has not immediately commented on the allegations in the report, while Taiwan's publicly released statement did not link the latest incident to Beijing.

Growing Concern Over AI-Enabled Hacking

The use of AI in cyber operations has become an increasing concern in recent years, with more advanced models potentially allowing attackers to identify targets, locate security vulnerabilities and assist with exploitation more rapidly.

Chris Thomas, a security researcher at code-security company Semgrep, said the Taiwan incident demonstrated how quickly AI-assisted hackers could potentially move toward their targets.

He cautioned, however, against viewing the technology as a completely autonomous hacking system.

Human operators, he said, are still involved in determining targets, setting objectives and directing AI systems.

“This is not fully autonomous. There was a skilled operator behind it,” Thomas said.

The incident highlights growing concerns among cybersecurity experts that increasingly capable AI systems could lower the technical barriers to sophisticated cyber operations and allow human attackers to conduct activities at greater speed and scale.

Taiwanese and Chinese authorities have not publicly provided further comment on the alleged China connection reported by the Financial Times.