Tempus’ USC partnership capped another week of AI-enabled clinical infrastructure growth, but a separate development highlighted the remaining evidence gap for healthcare AI outcomes. Researchers argued in Nature Medicine that many deployed clinical AI systems still lack rigorous assessments linking tool accuracy to improved patient health outcomes. The critique centered on ambient AI scribes and other hospital-deployed tools that evaluate or summarize clinical information, noting studies often measure clinician or patient satisfaction rather than whether AI changes clinical decision-making and downstream outcomes. For biotech and med-tech, the takeaway is operational: as more AI products enter routine workflows, regulators, buyers, and developers face pressure to demonstrate not only performance metrics but also measurable clinical impact.