You’ve probably read about ‘deep fakes’ and other ways of manipulating images. The problem with this trend is not so much the manipulation of the image, but the manipulation of the recipient. When AI, gan networks and other modern technology is used ‘for bad’, the impact will also be bad. However, it’s not the technology in itself that is the problem, but the implementation of it.
At JumpStory we use AI to make life easier for marketers and designers across the globale. Currently we have customers in +70 countries, and they take advantage of our different ways of using AI to complete relatively complex tasks …
A digital image is basically just an image composed of picture elements, namley pixels, each with finite, discrete quantities of numeric representation for its intensity or gray level that is an output from its two-dimensional functions fed as input by its spatial coordinates denoted with x, y on the x-axis and y-axis, respectively.
For a computer this means that you have to teach it to understand these billions of different compositions in order to be able to tag the images correctly. And that is one part of how JumpStory used AI to index and tag the millions of photos, videos, icons, vectors and illustrations that you can download on our platform.
Removing the background automatically using AI
Another feature on JumpStory is to remove the background on images. There are approximately 20 million more interesting activities than removing backgrounds by hand, and thanks to our clever AI, you can slash editing time – and have more fun! The technology used is called machine learning, by the way.
Matching your copywriting with relevant photos
A third part of the JumpStory platform is our ‘TextMatch’ feature, where we automatically help you match texts with images in seconds. The technology used is a combination of the machine learning mentioned above and then another part of AI – typically referred to as ‘Natural Language Processing’, where you teach the computer to understand and make sense out of words and sentences.
Predicting which images will perform best
A feature, which is still in beta-mode on JumpStory, is our “HighJumper”. We use a combination of neuro-marketing parameters and social media data to try to predict which images will perform the best. This is another great tool to help marketers choose between similar looking images to get the most impact from their marketing efforts.
Upcoming AI features on JumpStory
So how do we work with AI in the future at JumpStory? Is our goal to create thousands of fake faces as seen on other online projects, or what is our mindset, when it comes to the future of marketers, designers and AI?
Of course I can’t reveal everything, but some of the very exciting stuff that we’re working on at the moment includes:
Make photos look local every time
Use case: You have taken a nice photo with some models from the US, and you want to be able to save money and not have to make the same photo sessions in other parts of the world, where you also do business. With our tool you can make the photos look local every time.
Filter images by emotional response
Use case: A lot of times we’re not actually looking for an image, but for a particular image / response. We use AI to ‘tag’ and understand images based on the response that they have on the person looking at them. This is a complex combination of colors, composition, facial expressions and much more.
Turn still photos into short videos
Use case: You have some photos that you really like, but perhaps you also want to see them move a bit. You can use our future AI-tool to turn still images into small animations automatically.