As reported by BleepingComputer, a likely Russian-speaking threat actor orchestrated a global campaign using hundreds of AI agents to exploit two actively exploited PaperCut NG/MF vulnerabilities — CVE-2026-81578 and CVE-2026-82078 — compromising at least 440 PaperCut instances across 395 organizations in 48 countries. GreyNoise's threat intelligence reveals a campaign that compresses what traditionally took weeks into hours, and this is the detail defenders should fixate on.

Security Impact: As reported by BleepingComputer, a likely Russian-speaking threat actor orchestrated a global campaign using hundreds of AI agents to exploit two actively exploited PaperCut NG/MF vulnerabilities — CVE-2026-81578 and CVE-2026-82078 — compromising at least 440 PaperCut instances across 395 organizations in 48 countries.

Vulnerability Summary

CVEProductSeverityExploitation StatusPatch Availability
CVE-2026-81578PaperCut NG/MFCritical (actively exploited)Confirmed in the wildCheck vendor advisory
CVE-2026-82078PaperCut NG/MFCritical (actively exploited)Confirmed in the wildCheck vendor advisory

Both vulnerabilities were flagged as actively exploited earlier in September 2026. The attack chain observed by GreyNoise proceeds from PaperCut initial access through LSASS memory dumping, pass-the-hash, the noPac attack (leveraging unpatched CVE-2021-42278/CVE-2021-42287), and ultimately DCSync to extract full NTDS.DIT credential dumps. In one case, the adversary moved from initial access to full domain administrator in seven minutes.

Why the AI Angle Changes Everything

The use of OpenAI Codex and DeepSeek models to build, test, and refine exploits is not a novelty footnote — it is an inflection point. The attacker went from an empty workspace to first RCE against a live victim in under four hours, and to domain admin two hours later. Once the full campaign launched, 11 organizations were compromised in 26 seconds.

The threat isn't just that AI makes attackers faster. It's that AI collapses the gap between vulnerability disclosure and weaponized exploitation to a window that traditional patching SLAs cannot meet.

Historically, defenders had days to weeks between a proof-of-concept appearing and mass exploitation. AI-driven exploit development compresses that to hours. Organizations running patching cycles measured in 30-90 days are operating on a timescale that no longer matches the threat.

Who Is Most Exposed

Why the AI Angle Changes Everything
Education sector — accounted for roughly half of all breaches, consistent with PaperCut's heavy deployment in K-12 and higher education environments.
Organizations with internet-exposed PaperCut servers — the Netlas scanning platform was used for target enumeration, meaning any reachable instance was a candidate.
Environments running PaperCut on domain controllers or under domain admin service accounts — this configuration enabled direct domain admin escalation without intermediate steps.
Networks with unpatched noPac vulnerabilities (CVE-2021-42278/CVE-2021-42287) — the attacker actively chained these legacy flaws for lateral movement.

Immediate Actions

  • Patch PaperCut NG/MF immediately to the latest version addressing CVE-2026-81578 and CVE-2026-82078. If patching is delayed, disable or restrict external access to PaperCut web interfaces.
  • Audit PaperCut deployment architecture — ensure the service is NOT running on a domain controller and is NOT using a domain administrator service account. This is the single highest-impact configuration change you can make.
  • Verify noPac patches — confirm CVE-2021-42278 and CVE-2021-42287 are patched. These are years old but remain effective attack primitives.
  • Hunt for compromise — search for evidence of Mimikatz, Ligolo-ng, Certipy, BloodHound, Rubeus, Impacket, and NetExec execution. Look for LSASS access events (Event ID 4663 with access mask 0x1410), DCSync replication events (Event ID 4662 with replication-extended-rights), and unexpected account creations added to Domain Admins.
  • Rotate credentials for any PaperCut server that was internet-exposed since August 31, 2026, regardless of whether compromise is confirmed. Assume breach.

Shield53 Recommendations

Beyond immediate patching, this campaign demands strategic adjustments:

  • Adopt emergency patching SLAs for internet-facing critical infrastructure. The standard 30-day window is obsolete for exposed systems. Target 48-72 hours for critical CVEs with public exploit potential.
  • Implement network segmentation around print management systems. PaperCut servers should sit in an isolated VLAN with minimal AD trust relationships. They should never hold domain admin credentials.
  • Deploy credential guard and LSA protection on all PaperCut host systems to mitigate LSASS dumping effectiveness.
  • Monitor for AI-accelerated attack patterns — rapid, parallel exploitation attempts across multiple IPs targeting the same CVE are a signature of automated agent-driven campaigns. Geo-fencing and rate-limiting at the WAF can slow these attacks.
  • Reassess legacy vulnerability debt. The successful chaining of 2021-era noPac bugs demonstrates that old vulnerabilities never die — they become stepping stones. Prioritize patching based on exploit chain utility, not just CVSS in isolation.

The PaperCut campaign of September 2026 will not be the last AI-accelerated exploitation operation. It will likely not even be the most impactful. The organizations that survive the next wave will be those that treat patching velocity as a competitive security advantage rather than a routine IT task.