Google warns hackers are using AI to exploit security vulnerabilities.

(GLO) - Artificial intelligence (AI) is becoming a "double-edged sword" in the field of cybersecurity, as it not only supports defense but is also being exploited by hackers to detect and exploit dangerous security vulnerabilities.

The latest warnings from the Google Threat Intelligence Group (GTIG) security research team indicate that the AI ​​arms race in cyberspace has entered a more worrying phase.

According to Google, sophisticated hacking groups have begun using AI to search for and exploit zero-day vulnerabilities – flaws that have not yet been discovered or patched by developers. Notably, AI is no longer just playing a supporting role in writing code or analyzing data, but is directly involved in creating complete exploit code for cyberattacks.

AI is being used by hackers to find zero-day vulnerabilities faster. Photo: VOV

A shocking case announced by Google marks the first recorded instance of hackers using AI to generate zero-day exploit code outside of a real-world environment. The attack targeted a popular open-source code management tool and was able to bypass even two-factor authentication (2FA), which is considered a high level of security.

Security experts say the exploit code exhibits many characteristics of AI-generated content, such as an overly "standard" logical structure, lengthy comments resembling instruction manuals, and the presence of numerous data strings commonly found during AI model training. Google confidently asserts that AI was used to assist in the detection and exploitation of this security vulnerability.

Experts warn that the biggest risk lies in speed. Previously, hackers could spend weeks or months researching and building attack tools. But with AI, this time can be shortened to just a few hours or minutes. This is causing the "golden window" for businesses to patch vulnerabilities to shrink even further.

Beyond exploiting vulnerabilities, AI is also being used to create more sophisticated phishing campaigns. Google reports instances of impersonating technical support staff to steal Gmail user information using highly convincing AI-generated content.

However, AI is also being used by technology companies for cybersecurity defense. Mozilla says it used AI tools to detect and fix more than 400 security vulnerabilities in just one month.

Experts predict that in the future, the confrontation between defensive and offensive AI will become increasingly fierce, forcing organizations to accelerate patch updates, automate monitoring processes, and enhance their ability to detect cybersecurity threats early.

Source: https://baogialai.com.vn/google-canh-bao-hacker-dung-ai-tan-cong-lo-hong-bao-mat-post587647.html