Because the still image quality one gets from a (usually) tiny sensor with only a few megapixels can't compete with the still image quality one gets from a larger sensor with far more resolution.
HD video = 1,280 x 720 = 0.9MP
FHD video = 1920 x 1080 = 2.1MP
2K video = 2560 x 1440 = 3.7MP
4K video = 3840 x 2160 = 8.3MP
The sensor chips in many consumer grade video cameras are typically around 8.8 x 6.6 mm with 58mm² of area. To get sensors very much larger than that in a video camera, one must spend a LOT of money.
Current professional ILCs typically use sensors with anywhere from 20-50 megapixels or even more. And those sensors are usually at least APS-C (24x16 mm with 384mm²) if not FF (36x24 mm with 864mm²) or larger in size. The larger sensors allow less noisy and sharper images but they cost more. So do the lenses needed to cast an image circle that large at higher magnification to get the same angle of view with the larger sensor.
Because the fixed zoom lenses used by most video cameras are inferior to the lenses used by most professionals shooting with DSLRs (Digital Single Lens Reflex), MILCs (Mirrorless Interchangeable Lens Camera), or other ILCs (Interchangeable Lens Camera).
One can select a lens optimized for a particular task instead of being forced to use a jack-of-all-trades and master-of-none type of lens.
Frame grabs from video footage taken with even high end ILCs are relatively low quality compared to still frames from the same camera
This is because the processing power of the camera is usually limited and couldn't apply the same processing it applies to a still image to every video frame 24/30/60 times per second.¹ Not to mention that video compression schemes are totally different from those used for still images. Extend that down to video cameras with smaller sensors and more limited processing power and the image quality of a frame grab gets even worse. The quality of each single frame of video does not need to be the same quality as a still image to be perceived by our eye/brain vision systems as equally good when we're viewing the video at 24/30 frames per second. Our brains construct a better perception of the moving picture than what is contained in each frame of the video!
¹ This is beginning to become less and less significant as newer high end cameras can both shoot 8K video (which takes a lot more processing power than, say, FHD) and also take still images in bursts of around 30 fps which was unheard of just 3-4 years ago. But even then, video compression used to keep file sizes manageable usually means the quality of 8K frame grabs are lower quality than still images from the same sensor.
A big part of being a professional is being able to select the correct tool for the correct task. If one showed up to a still image gig with a fully rigged out RED Dragon (You can get the RED Dragon-X kit with a Sigma 18-35mm f/1.8 lens for around $20,000) with a sensor that is basically the same height as an APS-C still camera but a little wider for a 16:9 aspect ratio you might not be laughed out of the building like you would if you showed up with a consumer grade camcorder, but you'll still get more than a few raised eyebrows. You'd probably also not be considered for the next job they have. Why? Because you brought the wrong tool for the job.