SivaPrakash

The Scenario: Fixing Software Bugs (Vulnerability Management) In most IT teams, we prioritize fixes based on two things: Severity Scores (how "dangerous" a bug as per CVSS ) and Executive Pressure (what Security ops team is worried about). The Conflict: Imagine a "Level 10" bug exists on an internal server that nobody can access from the outside. Leadership wants it fixed immediately to stay compliant. However, an AI analyzing real-time web traffic predicts that a "Level 5" bug on your public login page is currently being targetted by hackers and will be attacked within 24 hours. The AI says: "Fix the small bug first." The Security ops team says: "Fix the big one first." How to Resolve the Conflict: To resolve this, we shouldn't just pick a side; 1. AI Transparency If the AI suggests a "weird" priority, it must provide the reasoning and evidence. If the AI can point to a specific data point (like a spike in "probing" traffic) that a human wouldn't notice, the team should follow the AI. If the AI’s logic is a "black box" that can't be explained, we stick to the human plan. 2. Plan based on "What’s the Worst That Could Happen?" We compare the "cost of being wrong" for both sides. If we ignore the AI: We risk a total system breach tomorrow. If we ignore Security ops team: We might fail a paperwork audit, which can be fixed later. Resolution: We prioritize the task with the highest immediate risk, which in this case is the AI’s prediction. Who Should Have the Final Say? The Human Team Lead (Also the Human in the loop or gets report from the human in the loop) should have the final say, and they must act like a judge. The Basis for the Decision: The human only overrules the AI if they have "Outside Information" that the AI doesn't know. Example: "The AI wants to fix that login page, but we are actually deleting that entire website tomorrow morning." The Rule: If the human has a fact the AI doesn't have, the human wins. If the human is just "going with their gut," the AI's data-driven sequence wins.