1
The image is enlarged
An upscaler creates a bigger pixel grid from the source. If you double both width and height, the result has four times as many pixels; that does not mean it contains four times as much genuine photographic information. Start from the original file rather than a screenshot or a copy that has been repeatedly compressed.
2
Missing detail is estimated
An AI upscale model predicts patterns that might belong between the original pixels. This can make hair, fabric, edges, and lettering appear clearer, but predictions can also change a person's features or invent texture. Treat the output as an interpretation of the source, not as evidence of what was actually there.
3
The result needs a visual check
Judge the image at the size at which people will see it, then zoom in to inspect faces, text, repeated patterns, and straight lines. Compare those areas with the original. If fine details look waxy, crunchy, or oddly regular, try a smaller enlargement or a less aggressive enhancement setting if one is available.