Google study on AI misuse in APT and malware workflows
Technical Analysis
Summary
Hide ▲
Show ▼
Google Threat Intelligence Group reported an unknown threat actor using PROMPTFLUX, an experimental VB Script malware, to query the Gemini API for just-in-time self-modification and obfuscation/evasion. Google said the sample uses a hard-coded API key, targets Gemini 1.5 Flash or later, and attempts persistence via the Windows Startup folder while also copying itself to removable drives and mapped network shares. The malware appears to be under development/testing and is not yet known to reliably compromise victims. Google also tied the report to broader Gemini abuse across China-, Iran-, and North Korea-linked activity for phishing, reconnaissance, C2 development, data exfiltration, and malware tooling.
Related Happenings
IPhone AI chatbot traffic leak of API keys, replayable tokens, and open relays
Technical Analysis
H score27
First: 30.06.2026 16:49
Last: 30.06.2026 16:49
Sources 1
About this happening:
**LLMKeyLens** testing found **444 iPhone AI chatbot apps** leaking **paid AI access**, exposing **API keys**, **replayable tokens**, and **open relays** that let others bill mode...
IPhone AI chatbot traffic leak of API keys, replayable tokens, and open relays
Technical AnalysisAbout this happening: **LLMKeyLens** testing found **444 iPhone AI chatbot apps** leaking **paid AI access**, exposing **API keys**, **replayable tokens**, and **open relays** that let others bill mode...
Google Cloud Vertex AI SDK Python predictable bucket squatting security flaw
Vulnerability
H score1
First: 16.06.2026 22:05
Last: 16.06.2026 22:05
Sources 1
About this happening:
**Google Cloud Vertex AI SDK for Python** had a **predictable temporary bucket** flaw that let an attacker hijack model uploads and reach **code execution** inside Google's servin...
Google Cloud Vertex AI SDK Python predictable bucket squatting security flaw
VulnerabilityAbout this happening: **Google Cloud Vertex AI SDK for Python** had a **predictable temporary bucket** flaw that let an attacker hijack model uploads and reach **code execution** inside Google's servin...
Google AI Threat Defense launch adds autonomous AI-attack detection and remediation for enterprises
Security Tool/Service
H score20
First: 28.05.2026 12:55
Last: 28.05.2026 12:55
Sources 1
About this happening:
Google Cloud launched **Google AI Threat Defense**, an **always-on autonomous** security platform aimed at stopping **AI-powered cyberattacks** across enterprise environments. The...
Google AI Threat Defense launch adds autonomous AI-attack detection and remediation for enterprises
Security Tool/ServiceAbout this happening: Google Cloud launched **Google AI Threat Defense**, an **always-on autonomous** security platform aimed at stopping **AI-powered cyberattacks** across enterprise environments. The...
Shai-Hulud worm clone activity on NPM
Malware Activity
H score69
First: 18.05.2026 12:45
Last: 18.05.2026 12:45
Sources 1
About this happening:
The **Shai-Hulud** malware activity has continued to evolve across the **npm supply chain** and related developer ecosystems. It first infected **npm packages** in **September 202...
Shai-Hulud worm clone activity on NPM
Malware ActivityAbout this happening: The **Shai-Hulud** malware activity has continued to evolve across the **npm supply chain** and related developer ecosystems. It first infected **npm packages** in **September 202...
Google GTIG analysis of adversary AI use for exploit development and attack orchestration
Technical Analysis
H score33
First: 11.05.2026 16:00
Last: 11.05.2026 16:00
Sources 1
About this happening:
**Google Threat Intelligence Group** published findings showing **adversaries using AI** for **exploit development** and **attack orchestration**, signaling that model-assisted tr...
Google GTIG analysis of adversary AI use for exploit development and attack orchestration
Technical AnalysisAbout this happening: **Google Threat Intelligence Group** published findings showing **adversaries using AI** for **exploit development** and **attack orchestration**, signaling that model-assisted tr...
Timeline
-
12.02.2026 14:45 3 articles · 4mo ago
Google publishes study on AI misuse by government-backed and criminal actors
Technical Analysis UpdateGoogle Threat Intelligence Group and Google DeepMind published findings on February 12, 2026 showing that government-backed cyber threat actors and financially motivated groups used AI across the attack lifecycle during the last quarter of 2025. The study described APT42 using generative AI to search for official email addresses and build reconnaissance on potential business partners, UNC2970 using Gemini to synthesize OSINT and profile high-value targets, TEMP.Hex using Gemini and other AI tools to compile target and separatist-organization intelligence, and APT31 using "expert cybersecurity personas" to automate vulnerability analysis and testing plans against US-based targets. Google also reported a rise in model extraction attempts, an underground jailbreak ecosystem, the Xanthorox toolkit advertising malware and phishing generation while relying on third-party and commercial AI products including Gemini, abuse of Gemini and OpenAI's ChatGPT public sharing features for ClickFix-style social engineering, and a September 2025 Honestcue case in which Gemini's API was used to dynamically generate and execute malicious C# code in memory.
Show sources
- Nation-State Hackers Embrace Gemini AI for Malicious Campaigns, Google Finds — www.infosecurity-magazine.com — 12.02.2026 14:45
- Nation-State Hackers Embrace Gemini AI for Malicious Campaigns, Google Finds — www.infosecurity-magazine.com — 12.02.2026 14:45
- Google Uncovers PROMPTFLUX Malware That Uses Gemini AI to Rewrite Its Code Hourly — thehackernews.com — 05.11.2025 17:33