Threat reportThreat IntelligenceTL-2026-4807
AI-Enabled Attack Trends: AI-Generated Malware, Social Engineering and Phishing (Bitdefender Business Insights)
AI-Enabled Attack Trends (TL-2026-4807), also tracked as The New AI Arms Race Starts Before the Attack, is a low-severity tracked intrusion set, first published 2026-10-30. It has no confirmed attribution, maps to 10 MITRE ATT&CK techniques (T1059.007, T1204.002, T1490), and is covered by 9 detection rules and 8 indicators of compromise.
- Severity
- LOWAssessed severity
- CVEs
- 0None referenced
- Techniques
- 10MITRE ATT&CK
- Actors
- 0Not attributed
- Detection rules
- 9SPL · KQL · Sigma
- IOCs
- 8Indicators of compromise
Key facts for TL-2026-4807
- Threat ID
- TL-2026-4807
- Also known as
- The New AI Arms Race Starts Before the Attack
- Severity
- LOW
- Status
- MONITORING
- Category
- THREAT_INTEL
- First published
- Last reviewed
- Attribution confidence
- LOW
- Motivation
- UNKNOWN
- Detection rules
- 9
- Indicators of compromise
- 8
Malware and tooling in AI-Enabled Attack Trends
Malware and tooling: Broomstick, FunkSec, Recipe Lister, Rhadamanthys, Claude Code
How AI-Enabled Attack Trends works
Bitdefender reports a 1,000%+ increase in AI-generated malware samples tracked by Bitdefender Labs (August 2025 to January 2026). In its survey of 1,200 IT/cybersecurity professionals, 59% experienced AI-involved social engineering, 56% encountered AI-generated malware attacks and 70% see increasingly sophisticated AI-enabled phishing. The article advocates a prevention-first strategy.
This is a threat-trend record, not a single exploit or campaign. It is based on a Bitdefender Business Insights article by Bruce Sussman (published 2026-09-09), 'The New AI Arms Race Starts Before the Attack'. The article cites Bitdefender Labs telemetry showing more than a 1,000% increase in AI-generated malware samples between August 2025 and January 2026. It also cites the Bitdefender 2026 Cybersecurity Assessment, a survey of 1,200 IT and cybersecurity professionals across six countries (published 2026-06-29). In that survey 59% reported AI-involved social engineering, 56% encountered AI-generated malware-based attacks, and 70% see increasingly sophisticated AI-enabled phishing. 53% believe AI helps attackers more than defenders. 44.8% report only partial visibility into AI use inside their organization, and implementing internal AI governance is named the top priority for 2026.
The article names no CVE, malware family, threat actor, C2 infrastructure or IOC, so this record contains no network indicators and no BeaconBeagle correlation was possible. The figures are vendor-reported. The underlying Bitdefender Labs methodology (how a sample is classified as 'AI-generated') is not published in the article and could not be independently verified.
For context, a Cloud Security Alliance research note (2026-08-31, 'The State of AI-Enabled Malware: Hype Versus Reality') summarizes Unit 42 analysis of 405 AI-associated samples. It reports that only 12 (3%) reached production endpoints and that existing behavioral analytics, sandbox detonation and code-signing anomaly detection blocked all twelve. The note names FunkSec ransomware, the Recipe Lister trojan, the Oyster (CleanBoost) backdoor, the Rhadamanthys stealer and a COM-hijacking component (360Util.dll) as examples. It also cites Anthropic's November 2025 disclosure of GTG-1002, a China state-linked actor that used a jailbroken Claude Code to execute an estimated 80-90% of tactical steps against about 30 organizations. These items come from a secondary summary and are included as corroborating context only; they are not described in the Bitdefender article.
Defender takeaways stated in the article: shift to a prevention-first posture, gain visibility into internal and unsanctioned (shadow) AI use, enforce web access, hardening and web filtering policies, build AI into the SOC or use an AI-leveraging MDR provider, and apply intelligence before an incident to identify exposure, reduce privileges and harden environments (continuous attack path removal).
MITRE ATT&CK techniques used in TL-2026-4807
Execution
T1059.007 Command and Scripting Interpreter: JavaScript; T1204.002 User Execution: Malicious File
Impact
Persistence
T1546.015 Event Triggered Execution: Component Object Model Hijacking
defense-impairment
T1553.002 Subvert Trust Controls: Code Signing; T1685 Disable or Modify Tools
Initial Access
Resource Development
T1587.001 Develop Capabilities: Malware; T1588.007 Obtain Capabilities: Artificial Intelligence
Defense Evasion
Remediation for AI-Enabled Attack Trends
Immediate actions
- Inventory internal and unsanctioned (shadow) AI tool use, including unauthorized cloud model access
- Enforce web access policies and web filtering to limit exposure to AI-enabled phishing and malware delivery
Workarounds
- Apply hardening policies to endpoints and environments to reduce the attack surface
Longer-term hardening
- Adopt a prevention-first security strategy rather than detection-only
- Implement internal AI governance, the top 2026 priority reported in the survey
- Build AI into the SOC or use an AI-leveraging MDR provider
- Apply pre-incident intelligence: identify exposure, reduce privileges, and continuously remove attack paths
Timeline of AI-Enabled Attack Trends
- Start of the Jan 1-6, 2025 window in which AI-assisted FunkSec ransomware samples were observed (per Unit 42 as cited by CSA)
- Bitdefender publishes its 2025 Cybersecurity Assessment survey of 1,200 IT/cybersecurity professionals
- Start (month granularity) of the August 2025 to January 2026 period in which Bitdefender Labs reports a 1,000%+ increase in AI-generated malware samples
- Anthropic discloses GTG-1002, a China state-linked actor using a jailbroken Claude Code for largely autonomous intrusion activity (as cited by CSA)
- End (month granularity) of the January 2026 Bitdefender Labs AI-generated malware measurement period
- Bitdefender publishes the 2026 Cybersecurity Assessment (1,200 respondents, 6 countries); page later modified 2026-07-13
- CSA research note 'The State of AI-Enabled Malware: Hype Versus Reality' reports only 12 of 405 AI-associated samples reached production endpoints
- Bitdefender Business Insights publishes 'The New AI Arms Race Starts Before the Attack' by Bruce Sussman
Sources cited for AI-Enabled Attack Trends
- The New AI Arms Race Starts Before the Attack (Bruce Sussman, Bitdefender)
- Bitdefender 2026 Cybersecurity Assessment
- Bitdefender: AI-Powered Security Across the Complete Attack Lifecycle
- CSA Research Note: The State of AI-Enabled Malware: Hype Versus Reality
- Unit 42: AI-Enabled Malware Analysis (as cited by CSA)
- Anthropic: Disrupting the first reported AI-orchestrated cyber espionage campaign (as cited by CSA)
- Global Cybersecurity Professionals Cite Top Threats, Concerns, and Challenges in New Bitdefender Report (2025 assessment)
Detection coverage for TL-2026-4807
As of 2026-10-30, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-4807 across Splunk SPL, Microsoft KQL and Sigma, covering 8 indicator(s) of compromise. The whole corpus is readable without an account; a free account unlocks full detection query text in Splunk SPL, Microsoft KQL and Sigma; paid tiers add raw indicator values, correlation and the MCP server. Threadlinqs MCP server · View plans.