Faced with the emergence of artificial intelligence, Emmanuel Mériot, Director France and Spain at Darktrace, the world leader in AI associated with cyber defence, warns: “Humans must erase themselves, it’s a battle between machines.”
To the extent that all the actors that make up the social fabric have adopted the Internet, cyberattacks are now listed alongside natural disasters and climate change in the annual list, drawn up by the World Economic Forum, of the most serious threats to our society.
Over the years,hackers have constantly reoriented their attack methods to dodge traditional defenses, forcing organizations to catch up and hastily find a Plan B in the presence of an attack..
Defences based on artificial intelligence have been a crucial response to the emergence of new and en increasingly sophisticated attacks. AI has been used to understand the “normal” state of a digital environment and to detect variations at the time they occur, signalling an abandonment of customary approaches to cyber defence.
A new phase in the evolution of attacks, hackers are now taking advantage of machine learning to deploy malicious algorithms capable of adapting, learning and improving to evade detection, indicating the next paradigm shift changement in the cybersecurity landscape: l’IA AI-based attacks. A recent study by Forrester found that 88% of security professionals expect a generalization of l’IA AI-based attacks. .
“Offensive AI” will harness AI’s learning and l’IA adaptability, ushering in a new era in which highly d’adaptation personalized, human-like attacks are scalableand spread at the speed of a machine. Offensive AI could hit a target’s network and use the information it sees to launch an attack, automatically including where valuable data is located..
“Thanks to its “aerial view” of digital activity, cyber AI will spot offensive AI once it begins to manipulate data”
New era.. We are already witnessing the first signs: the content of hypertrucage assistons (orou “deepfake”) manipulated by AI and designed to spread false information is a phenomenon of extreme concern for the giants of social networks. .
Relying on more data, defensive AI is based on unsupervised machine learning. It has a complete understanding of each user and each device in the entire network it protects, and uses this evolutionary understanding to detect variations that could be characteristic of an emerging attack. . With this “aerial view” of digital activity, cyber AI will spot offensive AI once it starts manipulating data..
Humans must erase themselves: it is a battle between machines. When this major breakthrough in attackers innovation will inevitably occur, the review, response and corrective actions will have to be achieved at speed and with the intuition of AI.. Only artificial intelligence can fight artificial intelligence.
We are only at the beginning of a new era of cyber, but we have reason to be optimistic: this is a new stage in the cyber warfare cybernétique that defenders have been preparing for for a long time, thus ensuring that when the starting gun of the l’IA AI-based arms race sounds, the “good guys” will get a head start.
Emmanuel Mériot is director of France and Spain at Darktrace.