tech
Your company's AI accounts are a hacking goldmine
Hackers are increasingly buying and reselling ChatGPT, Claude and Gemini credentials.
TL;DR
- Protecting and monitoring internal AI stacks is a major concern at the Black Hat cybersecurity conference.
- Hackers are stealing and reselling credentials for AI tools like ChatGPT, Claude, and Gemini.
- Attacks using AI have surged by 89%, with hackers leveraging AI to scale operations and target AI infrastructure.
- Companies are struggling with both sanctioned and unsanctioned AI deployments, creating new entry points for hackers.
- Using compromised AI accounts makes it harder to detect malicious activity and shifts token costs to corporations.
- There is a need for more advanced network monitoring tools to detect out-of-bounds AI agent behavior.
- Key recommendations include employee education, strong network policies against shadow AI, and automated network monitoring.