The difference is our brains can immediately judge the result of an image transform, but it takes a re-read to see if the text transform hasn't done something b

The difference is our brains can immediately judge the result of an image transform, but it takes a re-read to see if the text transform hasn't done something bad, like lost the meaning of the text. A blurry image is less likely to get you into trouble than saying the wrong words, so it is critical to validate manually what it says. That said, such tools could be useful for recommendations on how to rephrase things.