Block builds AI security workflows on OpenAI tech
What happened
On October 14, OpenAI News said Block uses AI systems including OpenAI Daybreak to speed up vulnerability discovery, detection and response, with humans verifying results. Its security workflows pick models per task and allow human takeover and agent cross-checks. In a recent red-team exercise, going from a high-level vulnerability description to an automatically proposed patch took 18 minutes; detection tuning replays seven months of production telemetry in under an hour. False-positive triage frees about 600 analyst hours per quarter and lifts triage throughput 50%.
Written by AI from the coverage · updated 2 hours ago
Coverage
Follow the reports to see the story from different sides.
- OpenAI NewsHow Block is building AI-powered security with OpenAI technology
Block 使用包括 OpenAI Daybreak 在内的 AI 系统加速漏洞发现、检测与响应，并由人工验证结果。在近期红队演练中，从提交高层次漏洞描述到自动提出补丁用时 18 分钟，检测调整方案可在不到 1 小时内回放七个月的生产遥测数据。其安全工作流按任务选择模型，支持人工接管和智能体互查，误报分流每季度释放约 600 分析员工时，分流吞吐量提升 50%。
Heat over time
Not enough continuous observations to draw a trend yet.