Since the link in your question is to one of Canon's regional marketing divisions, we'll assume you're using a Canon camera with Evaluative Metering. Canon tends to oversimplify its descriptions of how Evaluative Metering works. This video posted on Canon USA's YouTube channel is about as vague as the article from Canon Hong Kong. The user manuals are even less enlightening.
There are a myriad of factors that determine the weighting of each metering zone when using Evaluative Metering. As the EOS system has developed since it was introduced in 1987, Evaluative Metering has also evolved. It's gotten better with age as processing power and memory capacity have increased tremendously, not to mention the advent of color sensitive RGB+IR light meters.
Even back in the last decades of the ancient film days, primitive multi-segment light meters allowed for very basic analysis of which parts of the scene were brighter and which parts were darker. When semiconductors were introduced into cameras' logic systems in the 1970s, this information could be used to do rudimentary scene recognition. If the upper third of the frame was much brighter than the lower two-thirds of the frame, the camera would bias exposure on the assumption the photographer wanted the darker areas in the middle of the exposure range. If the upper two-thirds of the frame was brighter and the lower one-third was darker, the camera would bias exposure based on the assumption the photographer wished to expose the brighter areas in the middle of the exposure range. This worked fairly well for landscape photography, where the brighter area was sky and the darker area was the landscape.
Over the years light meters in SLRs and then DSLRs have multiplied the number of discrete segments from single digit numbers to dozens of them to hundreds of them to thousands of them. They have gone from being truly monochromatic to dual layer (measuring and comparing brightness at two different wavelengths of light) to today's RGB-IR sensors that are effectively miniature color imaging sensors. With the advent of mirrorless cameras (and Live View in DSLRs), metering can be done using information from the main imaging sensor.
As data rates and memory capacity available to camera designers have increased exponentially, the complexity of "library based" exposure metering routines has also increased. As a result, cameras are getting better and better at recognizing many different types of scenes and adjusting recommended exposure based on that identification.
Canon calls such scene recognition metering routines Evaluative Metering. Nikon does similar things with Matrix metering. Other brands have various names for it as well. There are a large number of patterns of light and dark areas, or even color information in cameras with RGB metering, to which the current patterns of brightness (and color w/RGB metering) measured by the various segments of the camera's light meter are compared.
The camera selects which is the best match from its library of preloaded patterns to the current pattern measured by the segments of the meter. It then calculates exposure based on the instructions for that specific pattern that have been programmed into the camera's firmware by the engineers who wrote it. It not only tries to identify which area of the scene the photographer wishes to expose in the middle of the range between too bright and too dark, but it also considers how bright the brightest areas in the scene are and how dark the darkest areas in the scene are and attempts to preserve as many of the bright highlights and as many of the dark shadows as it can.
So in a sense, using Evaluative Metering means the camera doesn't usually attempt to average the total brightness of the total frame with equal weight and base exposure on that. Rather, it attempts to recognize what sort of scene the camera has been pointed towards and expose appropriately, based on the decisions of the camera's developers, for the type of scene that it (hopefully) recognizes.
If it thinks you're taking pictures of clouds, it will be careful to not overexpose the brightest highlights in the scene. If it thinks you're taking pictures of a dark interior scene in a cathedral with narrow beams of bright light streaming through a few small windows, it will probably let the highlights go in order to preserve as much shadow detail as possible.
This is particularly the case when the photographer uses Evaluative Metering, in a situation where another metering mode, such as Partial or Spot would more accurately inform the camera exactly what part of the scene the photographer is most interested in exposing in the mid-tones between too bright and too dark. Or where the photographer limits blown highlights in any part of the scene by turning on Highlight Tone Priority (Canon)/Active D Lighting (Nikon)/Whatever other camera makers are calling it.
Is Evaluative metering exposing the scene to 18% gray?
Not intentionally, though I guess it's possible that under certain scenarios the exposure value calculation using Evaluative Metering could coincidentally come up with the same exposure value as averaging the entire frame to be exposed at 18% would be.
If you desire the camera to expose so that a certain part of the scene is 18%, then use Partial Metering or Spot Metering to meter that specific part of the scene. If the entire scene is uniform brightness, such as if a gray card fills the entire frame with equal brightness, Then Center Weighted Metering will also do the same thing.
But cameras still can't read the photographer's mind, even if they are getting better at guessing what the photographer probably wants.
Sometimes the camera can only be as smart as the photographer allows it to be.
This question covers how 18% metering, as well as how Evaluative/Matrix/Etc. library and/or algorithm based metering works.
How does 18% exposure algorithm work?