When Google added its search by image functions a month ago I didn’t really see much use for them. Why would I want to find an image that was similar to the one I already had? It was only a few days ago that I came across a possible use for the search by image.

Whilst purchasing a stock image for a client, I wondered if the same image had been used many times on other websites. Using Google’s search by image I decided to drag a watermarked copy of the image in from iStock, only to find that the search returned many high res un-watermarked versions of the same image. To see if this was a one off I decided to test with a completely different image and low and behold I was given full high res versions of that image also. Previously if you were to find a nice stock image that you wanted to obtain unlawfully you would have had to search for it using words to describe the image. This was obviously flawed as many web developers name their images in a non-descriptive way. What Google has created is a way of turning all stock image sites into a catalogue of easily obtainable images. At this point I must refer you to the disclaimer at the bottom of this post.

I’d be interested to see the long term effects of this on stock imagery companies. I for one would be coming up with a new way of making my watermarked images not so easily recognised by Google. The alternative would be to make them less easily downloadable but this would stop people being able to use comps in their mock ups.

Below is proof of how easily the above can be achieved and in this case would have been just as quick as purchasing the image legally.

Dragging the watermarked stock image into Google's image search

Search results containing all un-watermarked images

A high-res version of the stock image within 3 clicks

Disclaimer

I must state that I and elevatelocal do not condone the use of image search in this way. The image in the example above was purchased legally and the concept was merely an observation and not intended to be an instruction of any sort.