Memory usage in a modern OS is highly optimized. That is the OS tries to make the most efficient use of available memory. But efficient use of memory does not mean using less.
The old idea of memory use was that memory was a precious resource that should be used as sparingly as possible. That was necessary in old operating systems running on hardware with limited resources. That was the case even on mainframes through the 1950's.
But in the late 1950's the idea of a virtual memory OS was introduced. Note that virtual memory does not mean pagefile or swapfile. There is much more to it than that. This idea was further developed during the 1960's and by the end of the decade it was generally regarded as superior to other memory management systems. That remains true in 2022. This is for general purpose operating systems, not those for specialized needs.
Small computer systems, the kind that would be used by a home or small business, generally did not have the hardware resources needed for a modern virtual memory OS until the 1990's. By the dawn of the 21st century Windows, Linux, and the Mac all had a form of a virtual memory OS. This includes smartphones as well.
In such a system memory is recognized as the fastest storage medium and thus should be used to the fullest possible extent. Free memory is the ultimate in wasted memory memory, something that should be avoided if at all possible. The ideal would be zero free memory at all times, but we aren't there yet.
Memory usage in a modern OS is highly dynamic. The OS memory manager controls how much memory a process can have and how long it can keep it. The goal is always to optimize overall system performance, not to minimize memory usage. When there is no memory pressure processes are allowed pretty much whatever memory they want with few restrictions. This is best for performance.
But when memory pressure increases the OS makes some adjustments. The memory usage of processes is trimmed back to meet the increased demand. A modern OS also has several forms of cache whose size and contents are adjusted according to need and memory availability. All of this is done totally transparent to applications. The memory usage of an application on the same computer doing the same thing can vary by a factor of 10 or more depending on memory pressure. And all completely normal.
The computer in the screenshot appears to be operating normally as was designed. Unfortunately the principles used are not well understood by most computer users and that often leads to confusion.