Photo Credit: Getty Images
Hackers used an artificial intelligence system to conduct an extensive cyberattack against Taiwanese government agencies, raising concerns that autonomous technology could fundamentally change the nature of cyber warfare.
The attack, which unfolded over four days in July, reportedly involved AI agents mapping 21 government systems, compromising 85 government user accounts and extracting approximately 2,500 personnel records, according to Israeli AI company Dream, which discovered the activity.
Taiwanese authorities confirmed that the attack originated overseas and involved a combination of traditional hacking methods and AI agents. However, neither Taiwan nor Dream officially identified the perpetrators.
Experts suspect a connection to China, partly because simplified Chinese was found in internal documents associated with the attack. Taiwan has experienced persistent cyber pressure from China, which claims sovereignty over the island and has threatened to take it by force if necessary.
What made the latest incident particularly significant was the level of autonomy involved. AI was not simply being used by hackers to write code or identify vulnerabilities. Instead, the system reportedly coordinated as many as eight AI agents that carried out intrusions, assessed results and determined subsequent attack strategies without requiring constant human direction.
Dream described the system as operating much like a human hacking team. When one approach was blocked, the AI searched for alternative methods and adjusted its strategy.
Amir Becker, Dream's chief business and strategy officer, said the system represented a worrying development because it could strategize, learn and adapt independently.
The attack reportedly extended beyond government agencies to Taiwan's nuclear safety authority, government technology providers and at least seven companies in the energy sector.
Taiwan's Ministry of Digital Affairs said AI allowed attackers to combine different techniques and exploit weaknesses in secondary systems at greater speed and scale.
The incident has intensified concerns about the growing use of artificial intelligence in cyberattacks. While AI-assisted hacking has become increasingly common, experts say autonomous systems capable of managing complex attacks represent a significant escalation.
Kenny Huang, chairman of the Taiwan Network Information Center, described the incident as the first disclosed case of a fully automated attack against a government.
He warned that governments worldwide may not yet possess the legal frameworks, technical capabilities or policies required to defend against increasingly autonomous cyber threats.
Taiwan is already subjected to millions of cyberattacks each year. Government figures indicate that the country experienced an average of 2.6 million attacks from China every day last year, representing a six percent increase from 2024.
The latest incident therefore represents more than another cybersecurity breach. It signals a potential shift toward AI becoming an active participant in cyber conflict, challenging governments to develop defenses before autonomous attacks become more widespread.
