“Highlights AI failures in vulnerable situations using negative 'failed' framing, though the question suggests a path forward.”