Hackers Use Claude AI Agents to Automate Cyberattacks

Anthropic has reported that cybercriminals and state-linked groups are using Claude AI agents to automate cyberattacks, develop potential zero-day exploits and evade security detection. The activity was observed across espionage, financial crime and hacktivist campaigns.

Attackers used AI to automate phishing, reconnaissance, malware development, credential theft, data exfiltration and supply-chain attacks. In some cases, AI agents modified and rebuilt malware after it was detected, helping attackers bypass traditional security signatures.

The report highlights the growing need for advanced, behavior-based cybersecurity that can detect threats beyond known malware patterns. Organizations should strengthen endpoint monitoring, threat detection, incident response and data protection to defend against AI-powered attacks.

NPAV EDR help organizations detect suspicious behavior, correlate threats, monitor endpoints and respond to evolving AI-powered cyberattacks.