The X3 camera has a fixed aperture of f/2.8, and thus gives us only two controls for exposure: shutter speed and ISO sensitivity. That inflexibility all but demanded the use of ND filters to get appropriate shutter speeds (explanation below). The X5 camera gives us three controls for exposure: aperture, shutter speed, and ISO sensitivity. So in general, the X5 won't require ND filters as much as the X3. But they are still useful.
In video, the general goal is to use a shutter speed that is 1 over (2 x the frame rate). So if your frame rate is 30 frames/second, you want a shutter speed of 1/60 of a second.
This video will help explain why this is the case; the quick answer is "to get the correct amount of motion blur, so that the video looks good to the human eye and brain."
Another general goal is to use an ISO sensitivity as low as possible. The higher the ISO sensitivity, the more noise is introduced. I expect the lowest setting on the X5 camera to be ISO 100, which should give an excellent clean image.
So we start from there: ISO 100 at 1/60 of a second. And let's start with my favorite MFT lens, the
Olympus M.ZUIKO 12mm, which has a maximum aperture of f/2.0. On a bright day at f/2.0 and ISO 100, a shutter speed of 1/60 is going to be extremely overexposed. Even at f/2.8, a proper exposure might require a shutter speed of 1/1000 of a second. The question: how do we get that 1/1000 shutter speed, which produces jerky motion, down to 1/60, which produces smooth motion?
On the X3, we can't do it via the aperture (fixed at f/2.8), and we can't do it by lowering the ISO sensitivity (we are already at the mimimum of 100), so we have to do it with ND filters. Each "stop" on a filter effectively cuts the exposure in half. So a one-stop filter would allow us to change our shutter speed from 1/1000 to 1/500. A two-stop filter would allow us to lower it to 1/250, and a three-stop filter would allow us to lower it to 1/125. In this imaginary scenario, we would need a four-stop filter to get down to 1/60.
On the X5, however, we have control over the aperture. So instead of adding a 4-stop ND filter, we could instead close the aperture by three stops. That would mean changing it from f/2.8 -> f/4.0 (1 stop) -> f/5.6 (2 stops) -> f/8.0) (3 stops) -> f/11.0 (4 stops). That would give us an equivalent exposure, and allow a 1/60 shutter speed. As an added benefit, closing down the aperture increases the depth of field, which allows more of the scene to be in focus.
So why would we ever need an ND filter? Note that the maximum aperture on the 12mm lens is f/22. That seems like a high number, but it is only one more stop than we are already using (f/11.0). If it's an extremely bright day, then we might need 6 stops of exposure to get to our target shutter speed, and the lens only provides 5.
Further, lenses tend to perform the best in the middle of their aperture (e.g., f/8.0 here), and worst at the edges (f/2.0 and f/22 here). I would expect to see less corner sharpness and more vignetting at these extreme apertures.
The great thing about the DJI app is that it gives us a live histogram during video shooting, so we can see what the exposure is. Learning to read a histogram, and to adjust your exposure to get a good one, is probably the most basic and important skill that the X5 camera requires.