Kaspersky Threat Research has uncovered attacks on educational organisations in Latin America and Africa in which attackers used artificial intelligence to help direct activity inside compromised networks. Researchers traced reconnaissance, credential theft, movement between computers and data collection, including activity that compromised highly privileged administrator accounts.
The attackers used CyberStrikeAI, an open-source penetration-testing framework that combines a command-and-control system with a large language model. In this architecture, a human operator sets objectives, while the model can select tools, generate commands and analyse execution results. An implant installed on the compromised machine executes those commands and returns the results to the attackers’ server. Kaspersky observed the attacks between September and October 2026.
In one reconstructed incident, attackers initially worked manually through a compromised SQL Server, gathering information and preparing tools. After a two-week pause, they deployed the CyberStrikeAI implant. Activity progressed from implant deployment to the compromise of domain administrator accounts within three to four days. A final sequence of movement between computers, credential collection and database export took no more than an hour.
The AI operated on the attackers’ server rather than on the compromised computers. The implant functioned as a conventional remote-control tool, without making direct calls to AI services.
Kaspersky Endpoint Detection and Response Expert detected attack behaviours including credential dumping and malicious scheduled tasks. Kaspersky’s network detection tools identified CyberStrikeAI command-and-control traffic and suspicious queries used to map the victims’ domains. Kaspersky identifies these implants as HEUR:HackTool.Win64.CyberStrike.gen and HEUR:HackTool.Linux.CyberStrike.gen.
“The attackers used an AI agent based on CyberStrikeAI to carry out a full attack chain and adapt their next steps within compromised networks. However, the framework still relied on familiar techniques and procedures, including discovery, credential dumping, lateral movement, and collection and exfiltration. AI changes how attackers coordinate intrusions and the speed at which they operate, but well-configured security tools such as Kaspersky EDR can still detect and help stop the techniques they use,” said Nikita Nazarov, head of Threat Exploration at Kaspersky.
Researchers found Chinese-language comments in scripts, but these artifacts do not establish the attackers’ identity or location. Kaspersky could not attribute the activity to a known threat group or confirm that the same operator controlled all identified infrastructure.
Read the full research on Securelist.
To help defend against these attacks, Kaspersky recommends that organisations:
- Monitor behaviour on endpoints. Use an endpoint detection and response solution, such as Kaspersky Next EDR Expert, to investigate suspicious commands, credential access and persistence mechanisms, and respond to compromised hosts.
- Monitor activity across the network. Complement endpoint protection with network detection and response capabilities, such as those in Kaspersky Anti Targeted Attack, to identify suspicious traffic and investigate activity spanning multiple systems.
- Use threat intelligence to guide investigations. Give analysts access to current indicators and detection knowledge. Kaspersky Threat Intelligence Portal supports investigations into suspicious files and infrastructure, while its Hunt Hub provides threat-hunting and detection information.
- Maintain monitoring outside working hours. Where internal teams cannot provide continuous coverage, consider a solution such as Kaspersky Managed Detection and Response, which provides round-the-clock monitoring, investigation and response support.
