Alex Stamos says AI has pushed cybersecurity into a dark age

What happened
Alex Stamos, chief security officer at Cognition, stated that open-weight AI models now allow less experienced hackers to automate cyberattacks. He cited the use of AI-powered tools like ARTEX, which have been used in breaches of banks and other organizations. Stamos said that these tools can automate most steps of a cyberattack, making it easier and cheaper for attackers to find and exploit vulnerabilities. He warned that companies now need to test their own systems with AI tools within 90 days to avoid being breached first.
Key details
- ARTEX is a Chinese open-source attack tool used with a modified GLM-5.3 AI model.
- Attackers breached three South Korean banks and about 200 other companies in one week.
- A security investigation using a frontier AI model can cost $900 per incident.
- Microsoft's monthly Patch Tuesday has increased from 150 to over 900 fixes.
- Stamos joined Cognition as chief security officer in September.
Why it matters
Security teams at companies now face higher costs and greater complexity in defending against automated attacks, while attackers' costs are dropping. Organizations must quickly adopt new defensive measures, including using AI to test their own systems, or risk being breached. The rise in automated attacks affects banks, companies, and potentially any group relying on digital infrastructure, increasing the urgency and cost of cybersecurity efforts.
