Skip to content
Submit incident
AII-1000-65047
ReportedReported by fewer than three independent publishers, with no official finding.Supporting sources are pending verification.

DeepLoad malware uses ClickFix and AI-generated code to steal credentials

Security researchers at ReliaQuest identified a malware campaign called DeepLoad that uses the ClickFix social engineering technique combined with AI-generated code obfuscation to steal enterprise credentials and establish persistent network access. The malware hides functional code within large volumes of benign variable assignments designed to evade static security scanning tools.

30 March 2026 Β· Security and cyberattacks
30 March 2026
AII-1000-65047
Quick facts
Category
Security and cyberattacks
Country
United States
Date
30 March 2026
Drafted by
Raidu Incident Detection

What happened

DeepLoad malware combines the ClickFix social engineering technique with AI-generated evasion methods to steal enterprise credentials. The malware is delivered through links or files on compromised websites or SEO-poisoned search results, tricking users into executing malicious commands.

The payload hides inside legitimate Windows processes like LockAppHost.exe via asynchronous procedure call injection. The functional malicious code is buried beneath thousands of benign variable assignments, with researchers assessing that the volume and consistency of this obfuscation suggests AI-assisted code generation rather than human authorship.

Reported impact

The malware achieves rapid credential theft by scraping stored system passwords through a component called filemanager.exe, and intercepting live user sessions through a malicious browser extension. A hidden persistence mechanism using Windows Management Instrumentation allows the malware to reactivate itself even after attempted removal.

Index analysis

The large volume of padding and variable assignments in the code makes it difficult for file-based scanning tools to detect the malware. Researchers noted that the sheer scale of obfuscation likely rules out human authoring, and the use of AI suggests attackers could frequently alter the code to evade detection between waves.

Classification

Category
Security and cyberattacks
Industry
Digital security
Country
United States
Harm status
Actual harm
Intent
Intentional
Timing
Unknown
Harm types
Economic or property
Harmed
Business
Failure modes
Other
AI systems
DeepLoad malware

Sources and evidence

Sources not yet published

This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.

How sources are verifiedSuggest a source for this incident