Artificial intelligence already validates thousands of exploits autonomously to bypass two-step verification

Source Domain: www.apd.cat

Widespread Use of AI in Cyber Threats: Google has cautioned that several cyber threat groups are leveraging artificial intelligence on an industrial scale to augment attacks, evade detection, and automate operations previously reliant on human effort and extensive time.
AI Expanding Attack Volume: The significant shift is not just isolated incidents of AI usage, but its role in escalating attack volumes and sustaining prolonged campaigns.
Prevented Attack via AI Patch: Google applied a patch that likely curbed a planned attack exploiting a zero-day vulnerability in a Python script for system administration, which could bypass two-factor authentication.
AI in Offensive Cyber Tactics: The report highlights AI’s role in expanding exploit capabilities, code obfuscation, and avoiding security safeguards.
Case Studies of AI Use:
- Google’s APT45 example showcases the use of AI to validate thousands of exploits.
- The North Korean APT45 and Chinese UNC2814 use AI to disguise activities and study vulnerabilities.
AI in Mobile Backdoors: Promptspy, an AI-driven Android backdoor, moves autonomously to avoid uninstallation.
AI for Disinformation Campaigns: The pro-Russian Operation Overload used AI to clone voices for disinformation, and also used AI-generated tricks to hide malicious code from security systems.
Attacks on AI Components: TeamPCP’s attacks focus on stealing high-value secrets and AWS keys from AI software dependencies.