AI Accelerates the Search for Software Flaws on Both Sides
*Attackers are deploying AI to speed exploit creation, which is forcing defenders to rethink how they hunt and close vulnerabilities.*
The shift is already underway. Attackers now apply machine learning models to scan codebases and generate working exploits at higher volume than manual methods allowed. Defenders face the same pressure to locate the same flaws first.
Prior practice relied on human researchers reading source, running fuzzers, and reviewing commits. That workflow produced a steady but limited stream of reports. The new pattern introduces automated generation of candidate exploits, shortening the time between flaw introduction and potential use.
No public data yet shows which side currently holds an edge in discovery speed. The Wired account notes only that the pace has increased and that both offensive and defensive teams are adjusting their tools and staffing.
Why it matters
Teams that continue to treat vulnerability discovery as a purely human process will lose ground to automated systems. Organizations must decide whether to invest in their own AI-assisted scanning or accept higher risk of zero-day use by adversaries. The change is incremental rather than sudden, but the direction is consistent.
---
Sources:
No comments yet