Key Takeaways

  • 3AM’s European school campaign combines human direction with agentic automation across the attack chain
  • AI-enhanced phishing and autonomous malware can accelerate access, privilege escalation, and lateral movement
  • Schools can reduce exposure through identity controls, rapid containment, tested recovery plans, and executive oversight

3AM has targeted European schools with a ransomware campaign that blends human decision-making and an agentic AI loop. The approach reflects a broader shift in cybercrime: attackers can delegate repetitive intrusion activity to autonomous systems while retaining human control over targeting, data theft, and extortion.

This division of labor accelerates the attack lifecycle. An automated component can probe systems, process information, move laterally, and seek higher privileges without requiring an operator to approve each step. Human participants can then focus on the choices that carry the greatest financial value, such as identifying sensitive records, deciding what to exfiltrate, and applying pressure to affected institutions.

The campaign arrives in a particularly difficult environment for schools. ENISA identified ransomware as Europe’s most impactful cybercrime tool in its 2025 threat landscape. Education accounted for 1.7% of recorded incidents, a relatively small share that can still translate into substantial disruption for institutions with limited security staffing. Phishing remained the dominant intrusion vector.

Agentic automation does not need to invent an entirely new attack method to change the economics; it can simply make familiar methods faster and more scalable. The Microsoft Digital Defense Report 2025 found that AI-enhanced phishing produced a 54% click-through rate, compared with 12% for traditional phishing. That is 4.5 times higher, illustrating how automation can amplify the human-led initial-access phase.

The report also noted that autonomous malware can conduct lateral movement and privilege escalation without continuous human oversight. Applied to ransomware, that capability potentially shortens the interval between a compromised account and broader network exposure. How much investigation time remains when parts of the intrusion can proceed at machine speed?

The financial incentive remains clear. Between July 2024 and June 2025, financial motives drove 52% of observed global cyberattacks. Data theft or leakage appeared in 80% of investigated incidents. Encryption is therefore only part of the problem. Even when defenders prevent systems from being locked, attackers may already have copied student records, employee data, credentials, financial information, or internal communications.

There is some encouraging evidence. A 2025 survey by Sophos of 441 IT and security leaders across lower- and higher-education institutions found that 67% of ransomware attacks in lower education were stopped before encryption in 2025, up from 14% in 2024. That improvement suggests better detection and containment are having an effect. Still, stopping encryption does not establish that no data left the network.

For security leaders, the first priority is to constrain what a compromised identity can do. Phishing-resistant multifactor authentication, conditional access, least-privilege administration, network segmentation, and tighter controls around remote management can reduce an attacker’s room to maneuver. Logging should cover identity systems, endpoints, cloud applications, and unusual data transfers, not simply malware detections.

Tools such as Microsoft Sentinel and Security Copilot illustrate how defenders can use AI-assisted analysis to connect alerts and accelerate investigations. The advantage, however, depends on telemetry quality and established response procedures. An assistant working with fragmented logs or unclear escalation paths may only help analysts reach an incomplete answer more quickly.

Governance also belongs in the response. The NIST Cybersecurity Framework 2.0 gives education leaders a structure for identifying risk, protecting systems, detecting suspicious activity, responding to incidents, and recovering services. The EU NIS2 Directive adds pressure around risk management, incident reporting, supply-chain exposure, and executive accountability.

That said, schools do not need to chase every new AI security product. A more practical plan starts with rehearsed incident response, protected and tested backups, segmented administrative systems, vendor-access reviews, and clear reporting channels. 3AM’s hybrid model shows why those basics now need to operate at greater speed. Human attackers are still making the consequential decisions, but automation is giving them more opportunities to make those decisions sooner.