In brief
Intersection Observer observes the intersection of an element with a selected area. It is suitable for triggering work when approaching the content.
Useful to know
Intersection Observer gives asynchronous information about changes in the intersection of an observed element with a ancestor or with the visible area of the document. Instead of continuously calculating positions at each scrolling, you set up an observation and process the resulting records. [1]
A practical example is a marker after a list of results. When it crosses the selected area, the interface can request the next page. The task is to find the right moment to load, not to prove that the person has read all the previous entries.
It is important to prevent simultaneous repeat requests. Hold a flag for current loading and stop monitoring when there is no next page. These are application rules, not an automatic part of the observer.
Geometric crossing is not equivalent to user attention. Don't use it by itself as proof of readability or the quality of an advertisement's visibility.
Sources
For further detail and to check the information: