During its “Cloud Next” summit, the tech giant presented new AI cyber defense agents. The threat of large-scale automated cyber attacks allows publishers to sell new services, but it is also taken seriously by company security officials.
Do we need cyber attacker robots to defend against pirate robots? Google says yes. At its annual event “Cloud Next” in Las Vegas on April 22, the firm showcased a range of new cyber functionalities to defend against automated attacks. “AI brings new players into cybercrime, new techniques, and greater sophistication in attacks,” explained Francis deSouza, general manager of Google Cloud Security, in a call with journalists ahead of the event.
According to him, technology increases the scale of attacks and also accelerates their deployment time. He estimates that it now takes 22 seconds to move from one team of cyber attackers to another, compared to 8 hours before, such as during a transition between intrusion and espionage specialists.
To defend against this intensification and automation of attacks, Google believes there is no choice but to implement agents, themselves automated forms.
“We are moving from defense threatened by humans, to a defense where humans still intervene in the processes, to a defense completely guided by AI and supervised by humans,” said Francis deSouza.
David Grout, technical director for Europe at Mandiant (Google Cloud’s cyber subsidiary), adds, “Human teams alone can no longer absorb the volume of incidents: we must be able to defend at the scale of machines.”
The company therefore offers a fleet of AI agents consisting of a threat research specialist, another capable of sorting through detected threats, another for vulnerability detection, and one more capable of improving your defense strategy by analyzing data from third-party applications.
:fill(black)/2026/04/23/69ea6ebbae67a000429345.png "National tribute to the French soldier killed in Lebanon.")
