OpenAI 推动人工智能修复安全漏洞，而非仅仅发现它们

OpenAI is pushing artificial intelligence deeper into cybersecurity with a new focus: systems that can not only detect vulnerabilities but also help repair them. The shift moves AI security tools beyond scanning and alerting into the more difficult work of remediation, where developers must understand a flaw, assess risk, and apply a reliable fix without breaking other parts of the code. The ambition reflects a larger change in how companies are thinking about software security. Finding bugs is no longer the hardest part. Modern organizations already face thousands of alerts from scanners, bug bounty programs, audits, and automated testing systems. The larger bottleneck is turning those findings into safe, reviewed, production-ready patches. …