AI-Driven Malware Causes Real-World Harm and Evades Detection

The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.

Researchers have discovered multiple malware samples, including CLOSEDQUORUM and PROMPTFLUX, that use AI models for autonomous control, adaptive behavior, and self-rewriting to evade detection. These AI-powered malware have been used in live attacks, resulting in data and credential theft, highlighting a new wave of cyber threats.[AI generated]

Why's our monitor labelling this an incident or hazard?

The article explicitly details AI systems integrated into malware that have been used in live attacks causing harm, such as data exfiltration and device manipulation. The harms include violations of privacy, potential financial losses, and disruption to users' devices. The AI involvement is central to the malware's ability to evade detection and adapt dynamically, which directly contributes to the harm. Since harm has already occurred and the AI system's role is pivotal, this qualifies as an AI Incident rather than a hazard or complementary information.[AI generated]

AI principles

Privacy & data governanceRobustness & digital security

Industries

Digital security

Affected stakeholders

Business

Harm types

Economic/PropertyHuman or fundamental rights

AI system task:

Goal-driven organisationEvent/anomaly detection

AI malware is becoming its own category: Researchers found a malicious program that uses multiple AI models to decide what to do

2026-09-23

The Times of India

AI-Driven Malware Causes Real-World Harm and Evades Detection

Why's our monitor labelling this an incident or hazard?

Articles about this incident or hazard

AI malware can rewrite itself to evade detection

AI is helping hackers hide malware instructions on blockchains

A New Tool Found Malware That's Guided by an AI Hive Mind -- No Humans in Sight

Researchers Just Discovered New Malware That Lets a Panel of LLMs Decide How to Attack

Windows CLOSEDQUORUM malware uses AI models to autonomously select post-compromise actions

ClosedQuorum Malware Uses Gemini, DeepSeek, Qwen, and Mistral to Guide Attacks

Cisco Talos finds malware that puts its next move to a four-model vote

New malware lets commercial AI models call the shots: Talos

Introducing CAIRN: Frontier tracking for AI-integrated malware

Researchers Find Malware That Uses AI Models Instead of Human Hackers for Control - IT Security News

AI malware can rewrite itself to evade detection

Adaptive malware found: it changes behavior on the fly

A New Tool Found Malware That's Guided by an AI Hive Mind -- No Humans in Sight | Tech Biz Web

AI malware is becoming its own category: Researchers found a malicious program that uses multiple AI models to decide what to do

Researchers found malware that uses four different AI chatbots to run itself

CLOSEDQUORUM: New AI Malware Can Decide its Next Move Using Multiple Models

CLOSEDQUORUM: First Malware to Let AI Vote on C2

Cisco Talos Discloses Autonomous Windows Malware: Four AI Models Direct Each Attack

Windows Malware Turns AI Models Into Distributed Decision Makers

First 'autonomous AI C2 implant' uses panel of models to vote on next task

CLOSEDQUORUM, the malware that asks four AI models what to do next

Researchers found malware that takes its orders from a committee of four AIs, with no human

Four AI Chatbots Are Running a Malware Operation Without Human Commands

Cisco Talos descubrió un malware que consulta a cuatro IA para decidir sus ataques sin humanos

Un malware hace votar a cuatro modelos de IA para decidir cómo atacar una computadora: Cisco Talos

Así es ClosedQuorum, el malware para Windows que suplanta tus decisiones mediante un ataque con IA

CLOSEDQUORUM usa modelos de IA para decidir ataques contra Windows

CLOSEDQUORUM: el malware que vota con cuatro IA mientras tú sigues mirando solo al chatbot - TIC's en la Web