Translational risk is the chance that promising results in preclinical or early‑phase studies will not carry over to real‑world clinical benefit or safety. Non‑specialist decision‑makers should grasp this concept because AI tools may reduce but cannot eliminate translational risk; over‑confidence in model predictions can still lead to costly failures or unforeseen patient harms later.