How hacker breached more than 100 companies using AI agent — Report

A hacker has used artificial intelligence agents to break into websites belonging to more than 100 companies across several countries, raising fresh concerns about the growing role of AI in cybercrime.

A Reuters investigation published on October 19 reported that the operation targeted companies in the United States, Canada, Australia and Japan. Among the businesses affected were farm equipment manufacturer Deere, home appliance maker SharkNinja and engine manufacturer Cummins.

According to the report, the hacker relied heavily on artificial intelligence to carry out the intrusions, allowing automated agents to identify weaknesses and attempt to exploit them with limited human intervention.

Snorre Fagerland, a principal threat researcher at cybersecurity company NorForge, monitored the activity and described how the operation unfolded over several weeks.

The investigation found that the hacker used the open-source AI platform Cairn to direct the agents. The instructions reportedly required fewer than 100 Chinese characters, illustrating how relatively short commands could set complex automated processes in motion.

The agents spent hours probing targeted websites and attempting to exploit security weaknesses. Some successful intrusions reportedly took about 23 hours, while the estimated total cost of the campaign was approximately $5,000.

ALSO READ: Nigeria already a failed state, middle class has disappeared—Duke

That would put the average cost at around $50 per successful breach, although the figure is an estimate based on the reported operation rather than a general measure of the cost of cyberattacks.

The incident demonstrates how AI tools could lower the technical and financial barriers facing cybercriminals. Tasks that previously required substantial manual effort can increasingly be delegated to automated systems capable of running for extended periods.

However, the scale of an intrusion does not necessarily indicate the amount of damage caused. Reuters reported that affected companies confirmed breaches but said the damage or data compromise was limited.

The findings nevertheless highlight the risks organisations face when exposed systems can be identified and attacked automatically. Businesses may need to strengthen vulnerability management, monitor unusual activity and improve their ability to detect automated attempts to gain access.

The case also raises questions about responsibility when artificial intelligence is used to carry out complex operations. Although the technology can support legitimate security testing, the same capabilities can be redirected towards unauthorised access.

Cybersecurity specialists have warned that AI-assisted attacks may become faster and easier to scale as the tools improve. Companies, regulators and technology developers are therefore under pressure to ensure that security measures keep pace with the changing threat.

The investigation does not establish that every AI-assisted attack will achieve similar results. It does, however, provide a concrete example of how a single operator can use automated agents to target a large number of organisations.

The episode reinforces the importance of basic cybersecurity protections, including timely software updates, secure credentials, access controls and continuous monitoring of internet-facing systems.

(Reuters)