# Single Hacker Used ‘BOSSAI’ AI Agents to Breach Systems in 11 Countries

*Monday, November 9, 2026 at 10:06 AM UTC — Hamer Intelligence Services Desk*

**Published**: 2026-11-09T10:06:35.011Z (2h ago)
**Category**: cyber | **Region**: Global
**Importance**: 8/10
**Sources**: OSINT
**Permalink**: https://hamerintel.com/data/articles/22960.md
**Source**: https://hamerintel.com/summaries

---

**Deck**: Investigators say a hacker based in Taiwan used an AI‑driven “BOSSAI” setup with 12 cooperating agents to steal Taipei emergency dispatch records and penetrate hundreds of organisations across 11 countries, with confirmed data theft in three.

A case described by cybersecurity researchers shows how artificial intelligence is changing the scale of attacks a single person can mount.

According to a public technical report, a hacker in Taiwan built and used an AI‑based system called BOSSAI, made up of 12 software agents, to automate large parts of a broad intrusion campaign. The operator reportedly targeted hundreds of organisations in 11 countries and achieved confirmed data extraction in three.

Among the most sensitive targets was Taipei’s emergency dispatch system. The report says the hacker used BOSSAI to steal emergency dispatch records, exposing information on how the city coordinates responses.

The AI agents in BOSSAI were configured to divide up tasks such as reconnaissance, scanning, exploitation and data exfiltration. That division of labour meant the human operator could orchestrate a high volume of activity without manually handling each step against each victim.

For public‑safety agencies, the compromise of dispatch data is serious. These logs can reveal how fast services respond, where critical resources are based and which procedures are used during crises. For citizens, it raises the risk that details of calls for help and personal information now sit outside official systems.

The campaign also underlines how AI frameworks can make high‑end attack techniques more accessible. Instead of a team of specialists sharing duties, a single individual can direct multiple agents that each perform a piece of the intrusion process.

Defenders face a faster and more automated threat. Attacks that quietly probe many organisations across several countries can unfold before security teams have time to respond to the first signs of compromise.

The next concrete developments to watch will be whether law‑enforcement agencies in Taiwan and elsewhere publicly attribute the operation to a named suspect or group, whether emergency services in Taipei disclose the scope of affected data, and whether new incidents emerge that explicitly cite AI agent frameworks like BOSSAI as core tools in real‑world breaches.