One man in Taiwan used AI agents to scan hundreds of targets and steal data from about 10, a warning to anyone with old or unprotected systems online.
A single person in Taiwan used a set of AI helpers to hunt for weak spots at hundreds of organizations, and got into about 10 of them, researchers at Acronis's Threat Research Unit said last week. The haul includes emergency dispatch records covering seven Taipei districts.
The researchers are highly confident the attacker is one man based in Hualien City, Taiwan. He pointed AI agents, programs that carry out tasks on their own, at hundreds of organizations in at least 11 countries. Targets included emergency services, banks, universities and military systems. Data was collected or systems were accessed at about 10 of them in Taiwan, Mexico and Thailand.
The dispatch records came from a city's system for 119, Taiwan's emergency number for fire and ambulance services. They include incident times, response types, case status and street-level locations. They are case records, not live ambulance tracking. The agents also found a Taipei fire department platform exposing 2018 patient data, including ECG readings and ambulance video, with no login.
Commercial AI models mostly refused to help him attack. So the attacking agents ran almost entirely on "abliterated" models, which are AI models run locally with their safety training removed. The agents also made errors. One scanning tool invented flaws, and more than half of its high and critical findings were false positives or wrongly classified. The main attack agent reported results that had not happened.
Most important, the agents failed wherever logins were enforced, paywalls required a valid session, or endpoints were properly filtered.
What the agents found was mostly data that was already poorly protected. That included a medical data port left open since 2018, dispatch records on a public-facing API, a lender with 16 SQL injection points, and universities running unpatched Oracle databases. AI did not break strong defenses. It made it cheap to find and reach many weak targets at once. Organizations with basic protections in place have the least to worry about.
The alert box above lists the steps to take now. They center on knowing what you have online and locking it down.
The report does not say whether the affected organizations or Taiwan's authorities were told, or whether the exposed data can still be reached. His motive is also unclear. The researchers think he probably wanted money, and it is unclear whether he works for anyone. They tracked him through his own mistakes, including a reused identity across his businesses and attack servers, and his home IP address showing up in one of his chats.
AI did not beat strong defenses, but it finds forgotten systems with no login fast, so check what you have online and lock it down.
Malwarebytes now works inside Claude, so you can get a second opinion on a surprise text or link before you tap it, with no Malwarebytes account.
Read more: Malwarebytes tool lets Claude check suspicious texts and links for youA report describes browser traps that throw up fake virus warnings and feed impersonation scams in New York City. One panicked click or call can cost you money.
Read more: Fake Virus Pop-ups and Impersonators Are Trapping New York BrowsersIndia's Labour Ministry says scammers are posing as the National Career Service. The real service is free, so any request for money is a red flag.
Read more: India warns jobseekers: real government job service never charges fees
