Correct. However, research models like mice play a critical role in revealing key mechanisms of actions for many diseases. One can crudely think of it as a reverse engineering challenge - picking apart the key functions and program flow from obfuscated object code. Not all of the program logic or functions will have 1-1 correlates to humans, but these models can provide important clues for new therapeutics that might work in people. These types of experiments simply can’t ethically be done in humans.

And I entirely agree with your other point: university press releases are notorious for hyping research. It does the biomedical field a disservice, and gives the impression clinical trials are but an afterthought. Nothing could be further from the truth.