INSPECTING ARCHIVED INTELLIGENCE (OUTDATED VERSION).
Chinese group develops AI-powered attack tool
|
2026-11-17 18:50
CRITICAL
HIGH
AI-ENABLED ATTACK · AUTONOMOUS
EXPLOITED VULNERABILITY
VULNERABILITY DISCLOSURE
STATE-SPONSORED & ESPIONAGE
Executive Summary
AI-generated
A Chinese state-backed group is developing an automated attack tool for cyber espionage, according to Google's latest AI Threat Tracker report released on November 17, 2026. The attackers are leveraging AI tools for reconnaissance, vulnerability research and social media influence operations, with China being one of the few countries that have matched other nations like Russia, Iran, and North Korea in using AI for nefarious purposes. Chinese contractors are reportedly using large language models to translate and comb through tranches of stolen data, while others are experimenting heavily with tools such as Artex, which allows users to automate multiple phases of cyber intrusions at machine speed. Google has identified evidence of a new automated attack tool being developed by the Chinese government for likely use in cyber espionage operations, and threat actors have been using similar tools in real attacks or tested them in private.
Technical Mitigations AI-generated
• Apply the vendor fix for CVE-2026-109618, CVE-2026-109619 and treat internet-facing systems that were not patched in time as potentially compromised until verified.
Technical Observables 1 IoCs Login Required
Indicators of Compromise (IoCs) Protected
Raw threat telemetry, malicious IP addresses, file hashes, and direct VirusTotal correlation are restricted to authenticated users.
files 1
Intelligence Metadata
Actors / Malware / CVEs / Campaigns
CVE-2026-81578CVE-2026-81578
CVE-2026-40217CVE-2026-40217
CVE-2026-109619CVE-2026-109619
CVE-2026-82078CVE-2026-82078
CVE-2026-109618CVE-2026-109618
Target & Sectors
Global Scope
cryptocurrencycryptocurrency
financefinance
mediamedia
aviationaviation
aerospaceaerospace
governmentgovernment
energyenergy
defensedefense
educationeducation
healthhealth
technologytechnology
Incident Timeline
March 2026
The Chinese phishing service Jieyou added an LLM chatbot in June 2026, building on its existing model-agnostic chatbot functionality.
Click on any entity below to view its context and source!
tactic
Phishing
The Jieyou phishing service, active since March 2026, has a model-agnostic chatbot that allows clients to configure phishing panels and modify templates.
Separately, the Chinese
phishing
service Jieyou, active since March 2026, added an LLM chatbot in June 2026.
June 2026
The Chinese phishing service Jieyou added an LLM chatbot in June 2026.
Click on any entity below to view its context and source!
tactic
Phishing
Separately, the Chinese
phishing
service Jieyou, active since March 2026, added an LLM chatbot in June 2026.
2026/11/17
Threat actors are increasingly using AI-powered tools, such as Artex and PentAGI, to automate exploits, steal credentials, and scale cybercrime operations.
Click on any entity below to view its context and source!
organisation
Google
Chinese state group developing automated attack tool for cyber espionage, Google says.
North Korean-linked attackers used AI-generated personas to target cryptocurrency companies, with some of these personas containing Google's AI watermark, SynthID.
Artex has emerged as
the most used AI framework
for threat actors, mainly Chinese-speaking groups, but Google and Mandiant also observed attackers using and experimenting with open source frameworks such as Cairn, PentAGI, CyberStrikeAI, and HexStrike AI.
organisation
HexStrike AI
Google said it found evidence that Chinese government groups are now developing their own version of Artex for “likely use in cyber espionage operations.”
While Artex has garnered the most media coverage, Google said they have seen similar tools used in real attacks or tested by threat actors in private — including CAIRN, PentAGI, CyberStrikeAI and HexStrike AI.
Artex has emerged as
the most used AI framework
for threat actors, mainly Chinese-speaking groups, but Google and Mandiant also observed attackers using and experimenting with open source frameworks such as Cairn, PentAGI, CyberStrikeAI, and HexStrike AI.
Beyond the widely publicized ARTEX, these include CAIRN, CyberStrikeAI and PentAGI, along with experiments involving HexStrike AI and custom harnesses.
organisation
LLMJacking
In their latest AI Threat Tracker update, the researchers report that threat actors across cybercrime, espionage, and information operations (IO) are moving beyond simple prompting toward autonomous multi-agent frameworks, automated exploitation pipelines, and money-making schemes such as LLMJacking.
Attackers also use LLMJacking to make money.
The surge in
LLMJacking
has led to the proliferation of LLM Mixers, illicit marketplaces selling access to LLM services and underlying proxy infrastructure and routing traffic through pools of stolen credentials and compute instances.
organisation
Mandiant
Artex has emerged as
the most used AI framework
for threat actors, mainly Chinese-speaking groups, but Google and Mandiant also observed attackers using and experimenting with open source frameworks such as Cairn, PentAGI, CyberStrikeAI, and HexStrike AI.
Google Threat Intelligence Group (GTIG) and Mandiant say adversarial use of AI has reached a turning point.
Google Threat Intelligence Group (GTIG) and Mandiant have jointly released the 'AI Threat Tracker' report, which confirms that multi-agent frameworks are being used to launch attacks and automatically generate attack codes by comparing pre- and post-patch code.
threat_actor
APT42
State-sponsored groups known to have adopted LLMs include multiple Chinese APTs, such as Leaky Castle;
Iran-linked APT42
, also known as Calanque Ion; and North Korean hackers supporting IT worker activity and running attacks against cryptocurrency organizations, such as Midnight Neptune (
UNC1069
).
The Iranian actor
CALANQUE ION
, formerly
APT42
, uses them for reconnaissance and social engineering, and Chinese actors use them for research, scanning and exploits.
On the espionage side, Iran’s CALANQUE ION (APT42) uses Gemini for reconnaissance, and PRC contractor LEAKY CASTLE analyzes stolen data with LLMs.
organisation
PaperCut
As an example, Google cited two vulnerabilities that
emerged this summer affecting PaperCut
print management servers.
In late August, attackers chained PaperCut flaws CVE-2026-81578 and CVE-2026-82078.
The PaperCut print management server vulnerability is a leading example of automated AI-based attacks.
organisation
SOMBERMEME
A suspected DPRK cluster, MIDNIGHT NEPTUNE, created recruiter profiles carrying SynthID watermarks, then modified .claude/settings.json files to drop the SOMBERMEME backdoor.
The hackers modified code so that when users started up Claude Code, it would automatically trigger the download of a backdoor researchers call SOMBERMEME.
Fake recruiter profiles with SynthID watermarks led targets to
ClickFix
lures, and the attackers then edited .claude/settings.json in internal repositories so that starting Claude Code downloaded the SOMBERMEME backdoor.
organisation
SecOps
Google SecOps has released detection rules for ARTEX, and Mandiant has published a guide to help organizations secure their systems.
Google SecOps has released rule packs to detect ARTEX.
Google said its security operations (SecOps) team has deployed specialized behavioral and volumetric rulepacks to detect use of ARTEX.
organisation
ZeroHollow
“These observations additionally extend to the use of other tools and services, such as ZeroHollow, an inference service used for exploit development and LLM routing, and Agent Zero, an open source agent framework which features autonomous code execution and dynamic tool generation,” GTIG notes.
A PRC-nexus group was also building its own framework, and actors tested ZeroHollow and Agent Zero.
Others use ZeroHollow, an inference service for exploit development and routing, and Agent Zero, an open source agent framework that runs code on its own.
financial
$1 million
This type of activity can generate significant usage and billing increases, with cumulative resource abuse reaching $1 million in one case.
Single-day fraudulent charges topped $100,000, and in the worst cases cumulative abuse reached $1 million.
Mandiant handled incidents where fraudulent charges passed $100,000 in a day, and the worst reached $1 million before cleanup.
organisation
Influence
Influence and Espionage Threat Actors Go All In
GTIG assesses with high confidence that IO actors tied to Russia, the PRC and Iran, commercial Influence-as-a-Service vendors and inauthentic political monetization networks now routinely use LLMs.
According to Google, threat actors such as state-sponsored groups, commercial Influence-as-a-Service (IaaS) vendors, and inauthentic political monetization (IPM) networks are adopting LLMs for content generation and infrastructure design, which improves the quality of long-running influence campaigns.
organisation
AI Threat Tracker
In their latest AI Threat Tracker update, the researchers report that threat actors across cybercrime, espionage, and information operations (IO) are moving beyond simple prompting toward autonomous multi-agent frameworks, automated exploitation pipelines, and money-making schemes such as LLMJacking.
Google
released
its latest AI Threat Tracker report on Monday, writing that threat actors are rapidly evolving their use of artificial intelligence in attacks — now leaning more heavily on autonomous multi-agent offensive frameworks like Artex, CAIRN, CyberStrikeAI and PentAGI to execute multi-stage intrusions at greater operational velocity.
organisation
Groups
Groups focused on espionage are therefore likely to use them with caution, while cybercrime groups may accept a lower success rate in exchange for reaching more victims.
Groups are also experimenting with custom-built harnesses, it added.
organisation
JetNight
APIMart charges five times more once a prompt passes 100,000 tokens, while JetNight sells monthly subscriptions.
multiplies prices by five above 100,000 tokens, and JetNight sells monthly subscriptions.
financial
$100,000
Single-day fraudulent charges topped $100,000, and in the worst cases cumulative abuse reached $1 million.
Mandiant handled incidents where fraudulent charges passed $100,000 in a day, and the worst reached $1 million before cleanup.
financial
$0.10
Prices range from $0.10 to $25 per million tokens.
“This infrastructure routes buyer traffic through a pool of stolen credentials and compute instances to ensure reliable access, while also offering the ability to spread query volume across multiple victim environments to reduce the likelihood of detection.”
GTIG saw prices from $0.10 to $25 per million tokens on Russian- and Chinese-language markets.
financial
$25
Prices range from $0.10 to $25 per million tokens.
“This infrastructure routes buyer traffic through a pool of stolen credentials and compute instances to ensure reliable access, while also offering the ability to spread query volume across multiple victim environments to reduce the likelihood of detection.”
GTIG saw prices from $0.10 to $25 per million tokens on Russian- and Chinese-language markets.
organisation
Amazon
Amazon’s threat intelligence team similarly warned on Friday that it was observing an uptick in the use of Artex, citing campaigns aimed at stealing AI model credentials and bypassing identity verification.
Amazon reached the same conclusion after monitoring its own decoy systems: None of the techniques in these campaigns were new.
organisation
State
State-sponsored groups known to have adopted LLMs include multiple Chinese APTs, such as Leaky Castle;
Iran-linked APT42
, also known as Calanque Ion; and North Korean hackers supporting IT worker activity and running attacks against cryptocurrency organizations, such as Midnight Neptune (
UNC1069
).
organisation
Calanque Ion
State-sponsored groups known to have adopted LLMs include multiple Chinese APTs, such as Leaky Castle;
Iran-linked APT42
, also known as Calanque Ion; and North Korean hackers supporting IT worker activity and running attacks against cryptocurrency organizations, such as Midnight Neptune (
UNC1069
).
organisation
CALANQUE
The Iranian actor
CALANQUE ION
, formerly
APT42
, uses them for reconnaissance and social engineering, and Chinese actors use them for research, scanning and exploits.
organisation
Google’s Gemini
The report cites one instance of Iranian government-backed actor APT42 using several large language models, including Google’s Gemini, to conduct targeted reconnaissance and social engineering attacks.
organisation
Breaking News
“As the sophistication of these tools and underlying LLMs increases, we anticipate not only reduced human involvement in these operations but also increased stealth and sophistication of the activity.”
is a Breaking News Reporter at Recorded Future News.
organisation
Recorded Future News
“As the sophistication of these tools and underlying LLMs increases, we anticipate not only reduced human involvement in these operations but also increased stealth and sophistication of the activity.”
is a Breaking News Reporter at Recorded Future News.
organisation
TechRepublic
He previously covered cybersecurity at ZDNet and TechRepublic.
organisation
LLM Mixers
The surge in
LLMJacking
has led to the proliferation of LLM Mixers, illicit marketplaces selling access to LLM services and underlying proxy infrastructure and routing traffic through pools of stolen credentials and compute instances.
organisation
Impersonate
Related:
3 Charged in Scheme to Use AI to Impersonate CEOs and Attorneys to Steal $129 Million
financial
$129 Million
Related:
3 Charged in Scheme to Use AI to Impersonate CEOs and Attorneys to Steal $129 Million
organisation
Huntress
Two flaws were exploited as zero-days in late August, and Huntress and other researchers developed working exploits within 24 hours.
organisation
GTIG
GTIG expects that to change as the underlying models mature.
Tactical Metrics
Metrics
financial
$129M
Financial Impact
Click for context!
Related:
3 Charged in Scheme to Use AI to Impersonate CEOs and Attorneys to Steal $129 Million
Metrics
financial
$1M
Financial Impact
This type of activity can generate significant usage and billing increases, with cumulative resource abuse reaching $1 million in one case.
Single-day fraudulent charges topped $100,000, and in the worst cases cumulative abuse reached $1 million.
Mandiant handled incidents where fraudulent charges passed $100,000 in a day, and the worst reached $1 million before cleanup.
Metrics
financial
$100K
Fraud Losses
Single-day fraudulent charges topped $100,000, and in the worst cases cumulative abuse reached $1 million.
Mandiant handled incidents where fraudulent charges passed $100,000 in a day, and the worst reached $1 million before cleanup.
Metrics
financial
$0.1
Asking Price
Prices range from $0.10 to $25 per million tokens.
“This infrastructure routes buyer traffic through a pool of stolen credentials and compute instances to ensure reliable access, while also offering the ability to spread query volume across multiple victim environments to reduce the likelihood of detection.”
GTIG saw prices from $0.10 to $25 per million tokens on Russian- and Chinese-language markets.
Metrics
financial
$25
Asking Price
Prices range from $0.10 to $25 per million tokens.
“This infrastructure routes buyer traffic through a pool of stolen credentials and compute instances to ensure reliable access, while also offering the ability to spread query volume across multiple victim environments to reduce the likelihood of detection.”
GTIG saw prices from $0.10 to $25 per million tokens on Russian- and Chinese-language markets.
Intelligence Sources
Infosecurity-Magazine
2026-11-17
Fully Autonomous Agentic Attacks Are Three Months Away, Google Warns
Infosecurity-Magazine
The Cyber Express
2026-11-17
Google Finds Hackers Using AI Agents to Automate Cyberattacks
The Cyber Express
SecurityWeek
2026-11-17
Dailysecu
2026-11-17
TheRecord
2026-11-17
BleepingComputer
2026-11-16
Google warns that fully-automated AI attacks are coming soon
BleepingComputer
Security Affairs
2026-11-16
Cybersecurity Dive
2026-11-16
Datanet
2026-11-17
Unpublish from Social Media?
Are you sure you want to delete this podcast video from all synchronized social networks (YouTube, Facebook, Threads)?
Important:
Due to Meta API restrictions, Instagram Reels cannot be deleted automatically via API by third-party apps.
View Profile to Delete Manually
View Profile to Delete Manually
Tactical Intelligence
Report Intelligence Issue
Podcast Options
Generate
Incident Version History
CURRENT VERSION
Last Updated: 2026-11-18T07:15
Comprehensive Tactical Telemetry
Highly Correlated Entities
36x
organisation
Identified Entity
Google
entity
20x
timeline
Temporal Reference
2025
date
9x
tactic
Cyber Operation Type
Espionage
tactic
4x
attribution
Attributing Entity
FBI
authority
4x
tactic
MITRE ATT&CK Technique
T1566 - Phishing
technique
3x
source region
Origin Country
China
country
3x
general metric
Cases
8
cases
2x
source region
Origin Region
EUROPE
region
2x
financial
Financial Impact
$129M
2x
financial
Asking Price
$0.1
Contextual Telemetry
Context Block
11 METRICS
threat actor
APT Group
APT42
actor
general metric
Fake Accounts
100,000
fake accounts
target region
Target Country
France
country
general metric
Hours
24
hours
general metric
Gemini
4
gemini
general metric
Security Tools
100
security tools
general metric
Tools
150
tools
general metric
Sub
12
sub
financial
Fraud Losses
$100K
general metric
Early October
2,026
early october
general metric
Php Vulnerabilities
109,619
php vulnerabilities
Click on any entity below to view its context in the main text!
Selective Unpublish
Select the networks you want to remove this post from. The system will try to delete the real post through the API and clean the database so you can publish it again.
By navigating this website, you accept the use of strictly necessary technical cookies for session security and basic platform functionality. We do not use tracking or advertising cookies. Read our Privacy Policy.
Generating report…
0s