As damage from cyberattacks surges worldwide, cases have been reported in which AI carried out attacks autonomously, and it has been pointed out that AI does not overlook gaps in vulnerabilities. The number of information leakage cases is also said to be increasing in Japan.
Amid a sharp global increase in cyber-related damage, the US-based IT Resource Center (ITRC) announced that driver's licence information belonging to approximately 170 million people in the United States and Canada had been found on the dark web among leaked data.
In Japan, the number of cases announced in October as potentially involving information leakage reached approximately 52 million. At the car-sharing service Times Car, approximately 1.6 million driver's licence images were reportedly leaked.
Against the backdrop of advances in AI, DIVD announced for the first time an attack using agentic AI, explaining that it had demonstrated how AI autonomously determines its methods and moves on to its next action.
According to cybersecurity companies, attacks using AI, with humans intervening to provide instructions, have become realistic, and a breakthrough in which AI no longer overlooks remaining vulnerabilities began around July.
Microsoft announced that it had disabled more than 150 EvilTokens domains, reportedly helping to contain the spread of damage to organisations around the world.