Snapchat’s Memories and the Rise of the Visual-First Internet

Snapchat makes a clear statement that will continue to ring true as AI technologies improve social platforms: “The Internet is now a visual-first world, and we’re ready for it.”

Earlier this month, Snapchat introduced its new Memories feature, which allows users to save snaps and stories for future review and use. For a social platform that became popular for spontaneous, non-curated experiences, this move is surprising, with critics going so far as to claim that the change has now “ruined” Snapchat.

Speculation has pegged the decision as anything from a move into Facebook’s territory to an attempt to get more mileage from sponsored content. However, the more likely motivation is a basic recognition of the fundamental shift in the way we use smartphones, engage with content and communicate.

Visual content is on the rise

To understand what’s going on, we first need to look at the explosive growth of digital photography. Last year, more than 1 trillion photos were taken across phones and other devices. Social networks have played a huge role in encouraging that volume.

A recent report from Deloitte showed that there were more than 2,000 photo-sharing applications available by the end of 2015. Other findings show that social site users upload nearly 2 billion images per day onto social platforms like Facebook, Instagram, Flickr, WhatsApp and Snapchat. Cisco Systems analysts predict that 80 percent of global consumer internet consumption will be video content in 2019, and as recently as one month ago, Facebook’s Nicola Mendelsohn even intimated that text posts are declining on the platform in favor of video and images.

When internet use first became commonplace two decades ago, it was driven largely by the creation of search engines and the rise of Google. Text content, when crawled by web spiders, became understandable and part of a larger network. Internet usage as we know it now is moving from text-driven activities to image-driven activities. The challenge today is that visual information is much denser than text, full of information and nuance that is out of reach for traditional search.

The next step–engaging with photos

Think of a photo of friends in front of the Eiffel Tower. There are a lot of questions that can be asked about that single image. Who are these people? What’s in the background? Where was it taken? Is that the Eiffel Tower in Paris, or the replica at the Paris Casino in Las Vegas? The answers define how the image is interpreted and used in the larger world–and how we use our photos influences our relationships with others.

Which brings us back to Snapchat, the changes it has made and the efforts undertaken by other companies to make digital images more impactful for users. Facebook’s Moments app—which uses facial recognition to group photos based on who’s in them—encourages sharing between friends. Last month Pinterest previewed a visual identification tool that allows users to search for products based on the content of their images. Google Photos launched last year in an effort to help consumers better organize their photos through automated tools, and Apple recently announced the addition of similar functionality to its camera roll update.

These initiatives all encourage better integration between visual content and user experience. Social networks and tech giants like Google, Apple, Pinterest and Facebook have all implemented technology that helps automate processes through an understanding of the information locked within users’ photos.

Can AI unlock the potential of visual media?

If visual media is on the rise, and the information it holds could be a boon for search and other forms of digital engagement, it’s no surprise that there is a major push to find the key to unlock this potential. For many, the answer lies in artificial intelligence. Just as Google’s improved algorithms revolutionized text search, so too will new AI technologies make visual content fully accessible. These capabilities will completely change the way we interact with our images, both socially and informationally.

As anyone who has been incorrectly tagged in Facebook knows, there are limitations to current recognition technologies, but huge advances have been made in AI by linking major discoveries in neuroscience with computer science.

The next big breakthroughs will come through unsupervised learning–meaning that the AI can work on its own without continual guidance from computer scientists. Just as young children are able to explore and learn about the world on their own, so, too, will AI systems, and this will tear down the limitations that have held AI back until now.

Once unsupervised learning techniques make their way into consumers’ hands, the visual world will open up in ways that were only a dream just a few short years ago. Snapchat has already been exploring deep learning techniques for images and video. With the launch of Memories, Snapchat makes a clear statement that will continue to ring true as AI technologies improve social platforms: “The Internet is now a visual-first world, and we’re ready for it.”

Igal Raichelgauz is the co-founder and CEO of Cortica, creators of AI-driven image-recognition technology that is capable of human-level understanding of digital images.

IgalRaichelgauz