Started thinking about this after diving into some token mechanics—which got me wondering about something darker. When AI systems have power over living things, who's actually accountable when things go sideways?

Like, imagine this scenario played out with a human making the decisions instead. The blame would land somewhere concrete. But with AI calling the shots? It's murkier. If the algorithm messes up and something goes wrong, who takes the fall—the coder, the company, the AI itself?

Feels like we're building systems that can affect real outcomes but we haven't figured out the liability question yet. Curious what people think about this. Is this something we should be hammering out before these use cases become mainstream?