AI Brings Google Image Search into the 21st Century

Since 2000, Google has been working to make its search results more and more visual – including more imagery and videos in their searches, whether that’s illustrated weather reports, visual recipe experiences or live sports clips.

Google has done this by developing their ‘computer vision’, which is able to find the concept or meaning behind an image or video just by looking at its pixels rather than needing a description in words.

Further to Google’s computer vision, the company is making some fundamental changes to the way they think about search, using AI to make search experiences more visual, and therefore more enjoyable.

1. Immersive Visual Content

Google worked with the AMP Project to create AMP stories, which is an open source library that anyone can access in order to create a video story on the web.

Publishers have been experimenting with the ‘story’ format, which provides users with information in a much more visually exciting way. Found in Search and News, Google will also make these available in Google Images and Discover in the near future.

AI is also beginning to create AMP stories – finding the most interesting and pertinent information about a subject, and presenting it in a rich, and highly visual story format. Users simply tap on a story in order to be sent to the full article for more information, which is a new and exciting way for people to discover web content.

2. Visual Previews

When trying to gather information about a subject you do not know intimately, it can be difficult to filter the useful from the not-so-useful. Using computer vision, Google can understand the content of each video, and can therefore help its users to find the information that is most useful or relevant – showing these as ‘featured videos’.

3. Web Page Relevance

Google has made changes to Google Images in order to ensure that the webpage that the images relates to is high-quality and specific to the image itself. Having made changes to the Google Images algorithm, the results are now ranked on the quality of the image and the quality of the content on the page, with web authority becoming a more important signal for the ranking.

Google now prioritises fresher content, web pages where the image is central to the page’s content – or can be found high up on the page. Google will also start to show more context around the images in the search results, including captions that show the title of the webpage behind the image, and they are introducing suggested related search terms at the top of each Google Image search.

4. Google Lens Joins Google Images

Google Lens is already being used to identify items within a photograph on a phone, and it is now being added to Google Images as a way to explore, learn and shop from the visual search results.



Philip Likos-Corbett

Information Architect

The smartest, best looking and most captivating guy at Spot Studio, Philip also writes these biographies.