As opposed to photographing in colour, then having all the freedom to edit it afterwards?
Yes, if you shoot RAW. If you have difficulty visualizing an image in B&W, shooting in B&W gives you a good approximation of the final image at the time of shooting so you can adjust; many digital cameras can even process B&W with color filters, so if you have a particular type of processing in mind, such as using a red filter to darken skies (ala Ansel Adams's "Moon and Half Dome"), or a green filter to lighten foliage, you can get some feedback at the time of shooting to what it's going to look like.
And because you shoot RAW, you retain all the color information for B&W conversion if you change your mind in post. Only the JPEG preview is processed to B&W. So you can have the best of both worlds. The "B&W is irreversible" thinking only applies if you're shooting JPEGs.
Long exposure astrophotography is often done with a monochrome sensor to maximize the number of photons captured from a faint source. Relatively short exposures with separate Red, Green, and Blue filters are sufficient to color the image, but the longer unfiltered channel provides more detail in the structure of what's being imaged. You can replicate the same result with conventional sensor like in an SLR, but will need significantly longer total times because the color filters mean a lot more photons get rejected.
There is no benefit to shooting in a monochrome mode as the camera will just be taking a colour image and converting it to monochrome using built in settings. The closest you could argue as reasons to enable a monochrome mode are either to place an artificial inflexibility as part of a creative process or to view monochrome images in the camera display while shooting. Neither of those are in any way compelling options.
Conversely there are many benefits to shooting in colour such as being able to recreate the different tonal responses in films and potentially manipulate the image tonally before conversion.
There is one advantage with two big "Ifs".
If you are sure that what you do in-camera meets your artistic vision satisfactorily without the need to later shift the relation in grey tonal values between objects of different colors:
If you are saving the images as JEPG only files:
Then saving the jpegs in B&W will significantly reduce the file size of each image, all other things being equal. (Resolution, jpeg compression/quality setting. etc.)
I guess shooting color images is always the right option, 'reversibility' is the key. A color image could be turned B&W, but with a B&W shot, there's no option.
We all know that more or less the appeal of a photo lies in the post production, so why not shoot color and then decide.
Anyway, there won't be any difference in energy or information consumed by the sensors 'cause as far as I know they convert light onto digital codes and the color is just the result of filters used before it. You can say that a bright image uses more voltage than a dull one but color has no effect, none.
That depends on how much freedom you need/want.
Somebody using a large format view camera might be asking a DSLR user how it's even remotely possible to take half decent images with such a limited device.
Why would you use a telephoto lens giving up the freedom to crop a wide shot in post processing?
And why would anybody take images with a wide open aperture? For that blur you could photoshop into the image just the way you want?
For pretty much every property of a camera/lens/work flow/... you will find a certain value or range of values that you would like to have available for what you do. Other people might choose different values. If B/W is all you do without the desire to apply a lot of post processing, the freedom of color can become a burden.
There are even cameras that only take monochromatic images, like the Leica Monochrome (and video cameras like the Arri Alexa XT B+W or the RED Epic Monochrome, off-site as far as photo.SE goes, but on-topic as far as "spending 20k$ on something that has no point to it" goes, aka somebody buys this)
I find it helpful (at times) in composition. Color can be distracting both visually and emotionally. Other times, e.g. wildflower photography it helps to be able to see the image in color while arranging the various elements in the scene. I shoot RAW, and I have a camera (Sony a7R) with an electronic viewfinder.
In principle, yes, because unless the camera processing software is written in a very poor way, it will not first do the demosaicing to render a full color image and then use that image to compute the black and white image. Rather it will do a dedicated demosaicing to compute the black and white image directly from the raw data, the two processes are not equivalent, the latter will have less artifacts than the former.
The camera sensor has a color filter array, each pixel only records the grey values in either the red, green or blue part of the spectrum. The camera calculates the two other grey values for each pixel using interpolation (this is called demosaicing). Demosaicing is, of course, not magic; what the sensor didn't record cannot be recovered precisely from the neighboring pixels. So, you'll always end up with demosaicing errors.
If you want to have a black and white image then you need to recover less information using demosaicing compared to a color image, so this can be done more accurately. But if you first compile a color image and then compile a black and white image from that, the larger errors in the first step will propagate into the last step and you'll end up with a worse black and white image. While the larger errors will get reduced to some degree when you calculate a black and white image from a color image, you won't fully undo the larger errors.
The best thing you can do if you both want maximum flexibility to do post processing and the best quality black and white image, is to shoot raw images. You should then compile the black and white images yourself directly from the raw images by doing the required demosaicing yourself.