ESET Report Finds Malicious AI Skills and Adapting Threats
Summary
ESET's H1 2026 findings highlight malicious AI skills, PromptSpy's use of Gemini to adapt Android malware behavior, rising ClickFix and QR-code phishing, and more than 100 EDR-killer tools observed in the wild.
Key points
- ESET analyzed nearly 900,000 AI skills and found tens of thousands of suspicious instances and thousands classified as malicious.
- ESET identified PromptSpy as the first known Android malware to use generative AI in its execution flow; it uses Gemini to interpret interface elements and adapt across devices.
- ESET detections of ClickFix more than doubled between H2 2025 and H1 2026, with the technique appearing in AI-themed help pages, browser extensions, and cloud authentication scenarios.
- QR-code phishing reached record levels in ESET telemetry, with malicious links in QR codes steering users to interact on mobile devices.
- ESET Research documented more than 100 EDR-killer tools used in the wild, with new variants appearing regularly.
- Data from multiple sources showed a declining share of ransomware victims choosing to pay, which the article presents as possible progress in mitigation and response.
Article Details
- Publisher
- ESET (sponsored article published by BleepingComputer)
- Report Period
- H1 2026; ClickFix detections compared H2 2025 with H1 2026
- Scope
- Malicious AI skills, AI-enabled Android malware, ClickFix, QR code phishing, and ransomware-related EDR killers
- Sample Size
- ESET analyzed nearly 900,000 AI skills.
- Key Statistics
- ESET analyzed nearly 900,000 AI skills in H1 2026.
- ESET identified tens of thousands of suspicious AI skills and thousands of outright malicious instances.
- ESET detections of ClickFix more than doubled between H2 2025 and H1 2026.
- ESET Research documented over 100 EDR killers used in the wild.
- QR code phishing reached record levels in ESET telemetry.