Dec 13, 2024 (News On Japan) - A new AI system capable of quickly identifying vulnerabilities in corporate networks against cyberattacks has been developed.
Fujitsu announced the system on December 12th, explaining that it operates in a simulated virtual environment designed to mimic real corporate networks. Within this environment, AIs are divided into attackers and defenders to conduct repeated cyber battles. A separate AI then analyzes these interactions, identifies weaknesses in the network’s defense, and proposes countermeasures.
According to Fujitsu, traditional human-led countermeasures take weeks to months, during which time cyberattacks can evolve and become more sophisticated. In contrast, the newly developed AI can propose effective countermeasures in a much shorter timeframe.
Fujitsu has also released an AI tool designed to simulate customer harassment (known as "kasuhara") scenarios and provide training on effective responses.
Journalist: "Customer Service Center, how may I help you?" AI: "The remote control isn’t responding properly." Journalist: "We apologize for the inconvenience." AI: "Are you kidding me? If you handle this poorly again, there will be consequences!"
This AI tool has been trained on a vast database of real-world harassment incidents and corresponding responses collected from various companies.
Source: 日テレNEWS
