When you add an image to a post via the helper bar, you get this:
![enter image description here][1]
[1]: https://i.sstatic.net/ve8n0.png
What purpose does the enter image description here text serve? I've not noticed any way of seeing this text, except if I view the page source for the post.
The alt (alternative) attribute is used when the image fails to load, or for user agents that can't display images (screen readers, very old mobile browsers, etc.)
According to the HTML5 specifications,
The most general rule to consider when writing alternative text is the following: the intent is that replacing every image with the text of its alt attribute not change the meaning of the page.
This means that the "image description" hint is actually not very accurate. If the image is not important (in the sense that describing it will not help readers), is purely decorative, or if its content is already described by the surrounding text, then the attribute should actually be left blank.
Whatever you do, do not leave the placeholder text in, since that is worse than leaving it blank.
If you need a tooltip, you can use the title attribute. In Markdown, use
![alt text][1]
[1]: http://example.com/image "Your Title Here"
or
For example,
the same format is used for titles on links, like this.
alt text as a tooltip, since if you follow what the specs say the tooltip would just repeat the contents of the image. If you need a tooltip, that's what the title attribute is for.
The alt attribute is also used by people with eyesight problems who have difficulty seeing images so is not just a fallback.
So emphasises the text that @TimYiJiang gave from the HTML5 specification
The most general rule to consider when writing alternative text is the following: the intent is that replacing every image with the text of its alt attribute not change the meaning of the page.
In addition to what the others said, it's also good for searching purposes. Search engines like Google can pinpoint the subject of a picture much more accurately if there's a description in the alt text, than when they have to guess the subject from the context.
It adds the alt-text, which is basically a fallback if there is an issue fetching the image. If the image is not found, it'll show a white box with the alt-text in it.
On IE, the alt-text also gives a something similar to a title{*} tooltip, though this doesn't work on Chrome (or, IIRC, Firefox).
So it works exactly as it was intended.
*Curiously, webcomics still call it the "alt-text", even though they don't use the alt attribute. Remnants of a long-lost civilization when everyone used either IE or Netscape, I suppose :S
