Attackers are using AI agents to automate cybercrime at 'industrial scale,' FBI and AWS warn
Key Points:
- Attackers are leveraging the open-source AI penetration testing framework ARTEX to automate cyberattacks targeting banks, governments, energy, healthcare, and other critical sectors globally, according to FBI and Amazon Threat Intelligence alerts.
- Between October 3 and November 5, Amazon detected two ARTEX campaigns: one stealing AI credentials at scale to access models for free, and another generating deepfake videos to bypass live selfie identity verifications at cryptocurrency exchanges.
- The credential theft campaign used AI agents to harvest API keys and exploit various AI model gateways and application builders, while the deepfake campaign involved real-time video generation and forged ID photos to circumvent KYC checks.
- Both campaigns exploited poor security practices such as default credentials, exposed keys, unauthenticated interfaces, and unpatched software, highlighting the need for fundamental cybersecurity hygiene as the most effective defense.
- The FBI warned that ARTEX accelerates attack timelines, with some intrusions reaching privileged access in under three hours, and urged organizations to reduce external attack surfaces and implement zero-trust access policies to mitigate risks.