Google reported that Russian-speaking hackers created an AI agent that performed the first fully automated cyberattack in history.
According to the company’s data, the attackers targeted French websites in the media, entertainment, and education sectors.
The hackers, who used Russian in their technical documentation, configured the AI agents to continuously compare code before and after security updates were installed, detect vulnerabilities, and automatically develop exploits—without human intervention.
Researchers also discovered scripts designed to scout the internal infrastructure of compromised systems, crack passwords, and install web shells.
A web shell is a malicious script or small program that an attacker places on a web server to gain remote access and control over the internet.
One of the scripts was designed to remove web shells left behind by other hackers.
“Given the low barrier to entry, the ease of integrating new exploits, and the enormous potential for scaling, we believe it is highly likely that attackers will continue to use these frameworks in campaigns targeting organizations around the world,” the Google report states.
Previously, using AI models in cyberattacks required skilled hackers to direct their operations or correct errors.
This year, OpenAI’s autonomous AI agent also malfunctioned during an internal security test and attacked the infrastructure of the Hugging Face startup.
However, in that instance, the agent did not intend to cause targeted harm—unlike the attack reported by Google.
Підтримати нас можна через:
Приват: 5169 3351 0164 7408 PayPal - [email protected] Стати нашим патроном за лінком ⬇
Subscribe to our newsletter
or on ours Telegram
Thank you!!
You are subscribed to our newsletter