Cover art for FBI and AWS Warn of AI-Crime Automation

FBI and AWS Warn of AI-Crime Automation

Criminal groups use AI agents to breach banks and governments in hours, not days.

By ·

Nov 16, 2026

13 cards · 2m · tap to begin

From · · · 2m

FBI and AWS Warn of AI-Crime Automation

Criminal groups use AI agents to breach banks and governments in hours, not days.

In brief

Criminal groups use AI agents to breach banks and governments in hours, not days. Criminal groups deploy AI agents via ARTEX framework to automate cyberattacks on banks, governments, and critical infrastructure within hours. Originally reported by BigGo Finance.

FBI and AWS Alert

FBI and AWS Sound Alarm: AI Agents Now Automate Cyberattacks at Industrial Scale

The FBI and Amazon Threat Intelligence warn criminal groups use open-source ARTEX framework to run autonomous AI agents against banks, governments, energy and healthcare providers.

Two Campaigns Recorded

FBI and AWS Sound Alarm: AI Agents Now Automate Cyberattacks at Industrial Scale

Amazon's decoy servers captured two campaigns between October 3 and November 5: one harvesting AI provider API keys and another building real-time deepfakes for crypto exchanges.

Key Campaign Details

The first campaign extracted provider keys from exposed gateways like LiteLLM, One API and Dify, then pooled them for reuse. The second built deepfake pipelines to bypass identity checks.

Identity Fraud Tactics

Agents created real-time deepfakes at 11.5 frames per second, reverse-engineered liveness protocols, and forged identity documents matching real JPEG dimensions.

FBI Breach Examples

An August breach affected about 67,000 people in South Korea, and a September supply-chain attack impacted over 2,250 Microsoft tenants.

Weaknesses Exploited

Attackers targeted default credentials, exposed secrets, unpatched software, and interfaces without authentication or rate limiting.

China-Linked Groups

Amazon attributes moderate confidence that operators were separate China-linked groups based on simplified chinese prompts and china time-zone settings.

Cloudflare Findings

Cloudflare reported static CI/CD tokens with no expiration as a common root cause for intruders gaining access to victim accounts.

ARTEX Deployment

Internet scanning service Censys counted 1,465 hosts running exposed ARTEX consoles as of November 4, with likely higher numbers behind firewalls.

Security Recommendations

Organizations should replace default credentials, scan for embedded secrets, rotate exposed keys, and implement zero-trust access policies.

Detection Tools

Cisco Talos published indicator files tying 2,500 addresses to ARTEX and detection signatures for snort rules.

Microsoft's Findings

Microsoft reported attackers using agentic tools including ARTEX to hunt for GitHub secrets and API keys for AI inference.

Current Threat Level

The FBI states some intrusions reach privileged access in under three hours, significantly reducing the time to compromise systems.

The headline

Criminal groups deploy AI agents via ARTEX framework to automate cyberattacks on banks, governments, and critical infrastructure within hours.

Read the original on BigGo Finance

Read the TL;DR — this story on one card.

How we make summaries

React

Sign in to react and comment.

Comments (0)

Life is short. Keep it sweet. Respect others' opinions and be kind!

    Recommended next

    More decks on fbi and related topics.