The AI Attack Wave Has Arrived
November opened with a stark reminder that artificial intelligence has crossed the threshold from research curiosity to operational weapon. Three distinct attack campaigns disclosed this week demonstrate how threat actors are weaponizing AI across the attack lifecycle—from reconnaissance and social engineering to post-compromise data exfiltration.
The headlines tell a troubling story: attackers are no longer just experimenting with AI. They're deploying it at scale, and the sophistication gap between defenders and attackers is narrowing faster than many organizations realize.
VOLTA Infostealer: AI-Ranked Victims by Financial Value
Cybersecurity researchers have identified a new Windows infostealer called VOLTA-AUTODOXER (VOLTA) that represents a disturbing evolution in data theft operations. Unlike traditional infostealers that dump everything to command-and-control servers for manual review, VOLTA integrates large language models—specifically Claude, Gemini, and Kimi—to analyze stolen data in near real-time and rank victims by their monetary value.
The operational efficiency is chilling. VOLTA scrapes browser credentials, cryptocurrency wallets, banking information, and personal documents, then feeds this data to LLMs that assess financial worth, identify high-value targets, and prioritize victims for follow-on attacks. This allows operators to focus resources on the most lucrative compromises while ignoring low-value targets entirely.
Why this matters: Traditional detection relies on recognizing data exfiltration patterns. VOLTA's AI-driven triage happens locally or through legitimate API endpoints, making the malicious traffic blend with normal cloud AI usage. Organizations can no longer assume that stolen data will be manually reviewed weeks later—the assessment and exploitation window has collapsed to hours.
What to do:
- Monitor for unusual API calls to Claude, Gemini, and similar services from endpoints, especially if your organization doesn't use these platforms
- Implement application allowlisting to prevent unauthorized executables
- Review browser security policies and consider isolating credential storage
- Educate users that phishing payloads are now capable of instant financial profiling
SysGoRAT Targets Government Linux Infrastructure
A cyber espionage campaign targeting government Linux servers with SysGoRAT has been disclosed, notable for its use of suspected AI-generated scripts. The campaign specifically focuses on state-affiliated infrastructure, deploying remote access trojans alongside automation scripts that researchers believe were written or heavily modified by generative AI tools.
The use of AI to generate attack scripts offers threat actors several advantages: faster development cycles, code obfuscation through stylistic variation, and the ability to rapidly adapt exploits for different environments. While the scripts themselves aren't necessarily more sophisticated than human-written code, the velocity of iteration and deployment increases dramatically.
Why this matters: Government and critical infrastructure operators face adversaries with near-unlimited resources. AI-assisted development means these actors can maintain dozens of parallel campaigns, each tailored to specific targets, without proportionally increasing headcount. The traditional assumption that sophisticated attacks require large teams no longer holds.
What to do:
- Harden Linux servers with SELinux or AppArmor mandatory access controls
- Implement behavioral monitoring for anomalous process execution patterns
- Review and segment network access to state-affiliated or sensitive systems
- Maintain offline, immutable backups of critical system configurations
- Audit all remote access mechanisms and enforce certificate-based authentication
Microsoft Teams Phishing Exploits AD CS Misconfigurations
Perhaps the most immediately actionable disclosure involves attackers posing as IT support on Microsoft Teams to social engineer employees into granting remote access. Once inside, the attackers exploit insecure Active Directory Certificate Services (AD CS) configurations to obtain valid certificates, enabling persistent access and lateral movement.
The attack chain is elegant and difficult to detect: Teams allows external communication by default in many organizations, attackers impersonate legitimate support personnel during high-stress scenarios ("urgent security update required"), and once they obtain AD CS certificates, they possess credentials that appear entirely legitimate to security tools.
Why this matters: AD CS misconfigurations are widespread. Microsoft's own research suggests that the majority of enterprise AD deployments contain at least one exploitable certificate template. Combined with the trust users place in Teams as an internal communication platform, this attack vector bypasses most security awareness training.
What to do:
- Audit AD CS certificate templates immediately using tools like Certify or PSPKIAudit
- Disable external Teams communication unless specifically required for business operations
- Implement conditional access policies requiring MFA for any remote assistance tools
- Train users to verify IT support requests through out-of-band channels (phone, in-person)
- Monitor certificate issuance logs for anomalous requests, especially from user accounts
- Review the certificate template permissions and remove unnecessary enrollment rights
The Ivanti Problem Persists
As a brief but critical aside, Ivanti has patched two more Connect Secure zero-days (CVE-2026-111357 and CVE-2026-118527) that have been chained in attacks since at least October 1. Both vulnerabilities are now on CISA's Known Exploited Vulnerabilities catalog, meaning federal agencies and their contractors face mandatory patching deadlines.
Ivanti Connect Secure has become a favorite target for sophisticated threat actors, with multiple zero-day chains disclosed throughout 2025 and 2026. Organizations still running these appliances should treat them as compromised until proven otherwise.
Immediate action: If you operate Ivanti Connect Secure, patch immediately, perform forensic analysis for indicators of compromise, and consider migration to alternative remote access solutions given the platform's persistent security issues.
Synology NAS Under Active Ransomware Attack
Synology has issued an urgent warning about C3RB3R ransomware actively targeting network-attached storage devices. The surge in attacks suggests either a new vulnerability is being exploited or attackers have developed automated tools to identify and compromise internet-exposed NAS devices with weak credentials.
Small businesses frequently rely on NAS devices for backup and file sharing, often with minimal security hardening. C3RB3R operators understand this and are systematically scanning for vulnerable devices.
What to do:
- Update Synology DSM to the latest version immediately
- Remove NAS devices from direct internet exposure; use VPN access only
- Enable two-factor authentication on all administrator accounts
- Review and strengthen all user passwords
- Verify that backups exist on separate, offline media
The Pattern: AI Everywhere
The through-line in this week's disclosures is unmistakable: AI has become infrastructure for both attackers and defenders. VOLTA uses LLMs for victim triage. SysGoRAT campaigns deploy AI-generated scripts. Even the Teams phishing attacks, while not explicitly AI-driven, benefit from the same automation and scale that AI enables.
Defenders must respond with equal sophistication. This means behavioral analytics, anomaly detection, and security tools that can identify the patterns these AI-assisted attacks leave behind. It also means recognizing that traditional indicators of compromise—specific file hashes, known malicious IPs—are less useful when attackers can generate infinite variations.
The security perimeter has dissolved. The new battleground is identity, behavior, and the ability to distinguish legitimate activity from sophisticated mimicry. Organizations that haven't modernized their security architecture to account for this reality are operating with 2020 defenses against 2026 threats.
If your team needs help assessing exposure to these emerging attack vectors or hardening critical infrastructure against AI-assisted threats, Vici Tech Solutions offers penetration testing and security assessments designed specifically for the current threat landscape.