"But until these tools offer clearer and more widespread benefits for students, they should be evaluated with serious skepticism." This is the biggest takeaway

"But until these tools offer clearer and more widespread benefits for students, they should be evaluated with serious skepticism." This is the biggest takeaway I saw from this article. Specifically because measuring improvement in education is difficult. The most used metric for this is through the use of something known as a "learning gain". Typically, this is measured using a pre-test/post-test on a topic, with the intervention method in between. However, because an experiment needs to be strongly maintained, these experiments are often done on smaller scales, like for two to four weeks. Thus, the learning gains are only measured for a single topic. Semester-long or longitudinal studies are more difficult simply because it is harder to measure and control improvement over time while ensuring each student is given a chance to learn the material.