What are your thoughts on AI-driven vulnerability discovery platforms?

With the rise of AI technologies, platforms like 0day Rubbish are using AI to discover critical vulnerabilities. They utilize multi-LLM ensembles to identify 0-days, including a recent case study on an unauthenticated RCE chain in Cisco CUCM. How do you see the role of AI evolving in cybersecurity? Are there potential risks or ethical concerns we should consider?