Google Chrome now auto-generates image descriptions using machine learning


Google Chrome’s latest feature now auto-generates image descriptions using machine learning. The images are sent to Google to create the descriptions and users also get the ability to turn image descriptions on or off. This AI feature is aimed at helping blind and low-vision users.

A large portion of the content on the internet is visual, it becomes difficult for blind or low-vision readers to recognize the image if there is no presence of alt text. Now with Google Chrome’s new AI feature, when a user uses a screen reader, they can get descriptions of unlabeled images, for example, images that don’t have alt text.

However, it is important to note that the image descriptions are only in English, even if the browser is in a different language. Users can turn image descriptions on or off in Chrome Settings or in the context menu. Here’s how you can turn image descriptions on or off in the ext menu

  • Open Google Chrome browser
  • Open the context menu by pressing Shift+F10 in Windows, Ctrl + Alt + Shift + m or VoiceOver+Shift+ m  in Mac, Search+m in Chrome OS
  • Use the up or down arrow keys to select “Get Image Descriptions from Google”
  • Use the right or left arrow to open the image description menu.
  • It can be enabled for a single page or for all pages

Here’s how you can turn image descriptions on or off in Chrome settings

  • Open Google Chrome browser
  • Next to the address bar, select More More and then Settings.
  • At the bottom of the Settings page, click Advanced.
  • Under “Accessibility,” turn “Get image descriptions from Google” on or off.

The image descriptions are not perfect and if the algorithm isn’t able to recognize the image, it won’t try to label the image at all and the screen reader will say “No description available.” The new AI feature is only available to users with screen readers installed and enabled.

Commenting on this, Laura Allen, a senior program manager on the Chrome accessibility team said:

The unfortunate state right now is that there are still millions and millions of unlabeled images across the web. When you’re navigating with a screen reader or a Braille display, when you get to one of those images, you’ll actually just basically hear ‘image’ or ‘unlabeled graphic,’ or my favorite, a super long string of numbers which is the file name, which is just totally irrelevant.

Source 1, 2


Author: Manoj Nagendra

Manoj Nagendra is passionate about smartphones and the latest technology. He likes to write and explore the latest tech and you can often find him with an Android phone. You can follow him on Twitter @manojshesh24 and also mail at manojnagendra@fonearena.com