AI-enabled scams and ransomware operations drive recent security incidents

Mark Cuban reported that Facebook knowingly runs deepfake scam ads featuring his likeness for fake products. Ransomware affiliates and operators leveraged AI models to generate commands and used AI coding assistants for network reconnaissance and harvesting credentials. Developer platform WorkOS experienced a data breach after an ARTEX AI compromise of internal sales applications exposed contact records and communications. Mandiant responded to LLMJacking intrusions hijacking enterprise cloud and LLM resources. Threat actors used the ARTEX framework to scan GitHub for OAuth secrets, while others distributed fake webinar and event invitations featuring OpenAI, SpaceX, and Anthropic lures. China-linked groups used AI agents to scale espionage attacks. The week also had 6 disclosed vulnerabilities, down 36 from 42.