Ledger / Mar 2026
Dust Specter hits Iraqi officials with malware showing generative AI traces
Strong
Zscaler infers generative AI help from emojis and placeholders in decompiled code.
- Model
- Not namedLLM (model not named)
- Agent harness
- Nonechat or tool use only
- Attributed country
- Iranas attributed by Zscaler
- Sectors hit
- governmentIraqi government officials
Zscaler found a January 2026 campaign impersonating Iraq's foreign ministry whose malware carried emojis and placeholders suggesting generative AI helped build it.
“ThreatLabz observed several fingerprints in the codebase indicating that Dust Specter leveraged generative AI for malware development.”Zscaler
Sources
- Date
- Mar 2026
- Actor
- Dust Specter (nation state), as attributed by Zscaler
- Nexus
- Iran (nexus per Zscaler)
- Targets
- Iraqi government officials
- What the AI did
- Malware dev, Phishing content
- Model as named
- LLM (model not named)
- Model basis
- Inferred from artifacts
- Autonomy
- Assisted a human
- Outcome
- TWINTASK, TWINTALK and GHOSTFORM were dropped on targets.