Hot Take: too many image descriptions are ineffectively written and come off as inaccessible instead.

The main issue I've found is that people are giving too much information in the description. They're describing every last detail in the picture or video, whether or not it's actually relevant to the content.

This creates a wall of text - once fed to a screenreader, the person listening will just drown in words and will miss the actual pertinent part of the description. Just like how boring lectures can make you tune out!

For instance, "this is a screenshot of a FB post sharing a cookie recipe. The profile picture is of an Asian-passing femme wearing blue denim jeans and a tie dye tanktop. You can see specks of glitter in the left hand side of the picture."

Literally the only part of that image that's important is the cookie recipe. How the poster looks like is irrelevant. (Also potentially harmful if you're resorting to assuming race or gender.) It's just noise.

Or take memes. "An illustration of a dinosaur, possibly a Velociraptor, from the neck up, in a thinking pose. The background is a pinwheel of light green and dark green stripes. Text in white and a black outline reads "WHAT IF" up top and "THIS WAS A DREAM" at the bottom".

You could just say "the Philosophical Raptor meme, stating "What if this was a dream"". It wouldn't matter if the listener doesn't know what the meme looks like - whether it's blue polka dots or a thinking tarantula doesn't change the point; the joke is generally in the text.

The wordiness of most image descriptions is counterproductive, and I think it actually discourages people from making image descriptions. Writing all that detail is exhausting! Now try listening to it all!

But if we can keep image descriptions concise, I feel like that would make things waaaaaay more accessible for everybody.