Hackers are now using AI agents to Automate Cyberattacks – Report

Hackers are increasingly using autonomous AI tools to run multi-stage cybercriminal activities at a scale and speed that researchers admit bypass traditional human limitations, according to a new report by the Communications Authority of Kenya (CA), which warns that AI agents are now automating cyberattacks, operating around the clock, and targeting organizations faster than ever.

“Artificial Intelligence (AI) continued to influence the cyber threat landscape during the quarter, with bad actors increasingly using AI to support phishing, social engineering, credential harvesting, and other cybercrime activities,” reads in part the multi-agency report, sampled between July and September 2026.

CA’s new revelations further show that through the National KE-CIRT/CC, the agency detected over 2.6 billion cyber threat events during that period, representing a 14.45% increase from the threat events detected during the previous quarter.

It attributed the attacks largely to inadequate system patching, insufficient user awareness of phishing and other social engineering attacks, as well as the increasing exploitation of AI technologies by malicious actors to execute more sophisticated cyberattacks.

Similarly, web application attacks and system attacks remained the most prevalent cyber threat vectors during the period under review, according to CA, which it notes were primarily associated with the exploitation of vulnerable third-party libraries, inadequately secured Application Programming Interfaces (APIs), insecure serverless configurations, and weaknesses within open-source software dependencies.

“System attacks, on the other hand, were largely attributed to credential-stealing malware, weak access controls, information leakage, continued use of outdated and vulnerable operating systems, and user interaction with malicious links, among other attack techniques,” it adds.

The latest report by the Communications Authority of Kenya comes in the wake of a global trend in which recent reports showed that Chinese state-linked hackers used artificial intelligence, including the popular DeepSeek model, to launch cyberattacks against more than 200 organizations around the world in just 11 days, according to the cybersecurity company CrowdStrike, in a campaign that illustrates how AI is allowing hackers to operate on a scale that previously required far more human effort.

CA’s Director -General David Mugonyi says the authority remains committed to strengthening the capacity of stakeholders across the cybersecurity ecosystem to prevent, detect and respond to evolving cyber threats.

“Through targeted capacity-building programmes, adoption of relevant industry best practices and enhanced strategic partnerships, we will continue to strengthen our national cyber resilience and support the development of a secure and trusted digital ecosystem,” said Mugonyi.