SoftBank uncovers a large number of vulnerabilities with AI, re-diagnoses over 35,000

SoftBank president Miyakawa, at the earnings briefing, revealed that re-diagnosis of the company’s systems using GPT-6 Astra identified more than 35,000 additional vulnerabilities. The example is attracting attention as a model of AI-enhanced cybersecurity.

Why it mattersIt serves as a practical example of AI-based vulnerability detection and provides material for considering security-diagnosis efficiency improvements.