Cyberattacks have increased in recent months and the situation is unlikely to improve given Google’s statements. The Mountain View firm estimates that autonomous attacks should increase in the coming months.
How is this possible?
In recent weeks, the news has been punctuated by cyberattacks on government services. The French are therefore worried about the conservation of their personal data, but the situation could get worse.Google researchers discovered AI capable of carrying out cyberattacks autonomously.It can therefore search for vulnerabilities, prepare attacks and launch them without human intervention.
The report titled “AI Agent Threats Cross Important Threshold” was released on November 16 by the Google Threat Intelligence Group. Hackers have therefore gained skills thanks to AI that takes care of several steps. Whether searching for the flaw or attacking, it manages itself and does not need the intervention of a person. Several sites are targeted in France, especially in education, entertainment and media.
France is not spared
This phenomenon does not seem localized since it concerns several countries, including France. Certain French sites have even been particularly targeted by hackers. This whole story started with the discovery of a malicious server. The pirates had apparently forgotten to protect itor even delete it. By digging around, the researchers then found an AI-based hacking tool.
This system was made up of several AI agents capable of working together. The server was not empty since it contained the data of several victims.
How does the AI model work?
The cybercriminals, who could be Russian-speaking, would not have developed a complex system. Indeed, the tool is relatively simple since it “compares the code before and after the announcement of patches, analyzes the results to identify potential vulnerabilities, writes exploits and executes them automatically without human intervention”.
If previously, malicious people had to code to attack systems, this is no longer always the case. AI can also find flaws and implement a large part of the cyberattack. GTIG Chief Analyst John Hultquist explains that these hackers are exploiting technologies that do not require zero-day vulnerabilities.
They actually use the work of researchers to spot flaws. In addition, several groups of hackers have already used AI to carry out attacks. It will then be necessary to redouble efforts to secure systems and deploy patches.

