GitHub Reports AI‑Driven Code Pushes Demand Faster Secret Protection

GitHub’s latest blog post notes that one in three pull requests now involve an AI agent, up from less than one in ten a year ago. If the trend continues, most code pushed in the next two years could be AI‑generated and may never be fully reviewed by a human. To keep pace, GitHub and Microsoft Applied Sciences have released a fine‑tuned classifier that scans push candidates in under two milliseconds, potentially doubling the number of secrets caught before they reach public repositories. The data also shows that while the volume of screened pushes and credential‑carrying pushes has risen, the per‑push prevalence of secrets has not increased, and developers are overriding fewer push‑path blocks, indicating heightened caution. These findings underscore the need for automated, scalable secret‑protection tools as AI‑assisted coding accelerates.