Symbolic imageTaiwan confirms AI-agent attack on government systems
Taiwan's digital ministry said on Thursday that hackers used AI agents from overseas against government agencies in July, without detailing scale, damage or targets.
The Ministry of Digital Affairs said its monitoring units detected the "abnormal attack", which began on 20 July, and that the National Institute of Cyber Security issued alerts while it investigated. Attackers mixed manual work with AI agents such as OpenClaw, which chain attack techniques and use backup and test systems as entry points. Taipei named no country.
FT: eight agents, 21 systems
The Financial Times reported on Wednesday that suspected China-linked hackers were behind what it called a first-of-a-kind breach. It cited the Israeli company Dream, which said a tool built from open-source AI agents deployed up to eight of them across 21 government systems, compromised at least 85 accounts and extracted more than 2,500 personnel records before expanding to the nuclear safety agency and at least seven energy companies.
Attribution disputed
Dream named no group but said Simplified Chinese in internal communications made a China link probable; Beijing's Taiwan Affairs Office did not comment.
Labs concede models hacked systems
Al Jazeera reports that AI companies have said over the past fortnight that some of their own models managed to hack systems.
Other opinions
Al Jazeera covers only the labs' admissions; Le Monde adds Beijing's blanket denial of hacking.