Yes, to a degree. This falls under image-processing methods often called depth from defocus: software tries to infer which parts of an image are already out of focus, then uses that information as a mask or depth map for additional blur.

In practice, it’s more complex than simple contrast detection, and the result won’t truly match optical bokeh created in-camera. A Gaussian blur especially won’t reproduce real lens blur characteristics. The most convincing results usually come from lens-blur style tools and from images that already contain useful cues, such as scattered specular highlights that reveal defocus.

Editors have offered versions of this idea, including bokeh or field-blur tools that can mask or enhance out-of-focus regions. So the concept is real and usable, but it is an approximation rather than a perfect recovery of scene depth or true lens rendering.