When Algorithms Decide Who Leaves: Trust Crisis Behind Layoffs

Years ago, I led a team where we needed to enforce a forced distribution for annual performance reviews. At the time, our HR system had an automation tool that could generate an "efficiency score" based on metrics like code commit volume, bug fix rates, and work hour statistics. I tried it out and found that those ranking high were often colleagues who coded fast but with lower quality, while our core architect scored lower because he spent a lot of time on design reviews and documentation. I decisively turned off that feature and switched to manual calibration. Later, I told my team: Tools can assist judgment, but they cannot make decisions for us, especially those that change a person's fate.