The way we create and consume visual content has changed significantly with the advent of both digital and mobile cameras. Today, algorithms bring on the next wave of innovation and evolution in the space.
AI technologies are increasingly good at classifying, organizing and understanding images, changing the way both brands and creators think about visuals.
Sourcing visuals at scale
Cheaper and faster hardware, along with an abundance of rich data sets, has led to the visual computing revolution this past decade. Technology alters the entire creative process: ideation, shooting, editing, distributing, and discovery. The gap between mobile cameras and professional gear is quickly decreasing, giving everyone the ability to create photos and get their photos discovered. Pictures taken on iPhones are now used on the covers of magazines and in ads, so quality is no longer a barrier — and mobile cameras get better with every release. In a world where each photograph is competing for attention, how do we ensure that we don’t miss out on great content buried among noise?
This task is moving beyond the human scope of effort. A person simply cannot sort through the countless photos uploaded each day. At the stock photo service EyeEm, we’re using AI to help solve this problem. Our researchers and curators trained our algorithm to understand which photos are commercially viable and which aren’t. The only way to handle the huge volumes of content while ensuring that everyone is able to contribute is through technology.
AI can also be helpful on the creator side. An algorithm can analyze the photos on your camera roll and suggest the best to upload. It might sometimes feel intimidating to upload your photos to a platform full of beautiful imagery. An algorithm like this can help reduce that friction and encourage photographers to upload their photos by surfacing and recommending great photos they’ve taken. This AI can run offline and on-device (meaning nothing gets uploaded to the servers beforehand).
Sort by relevance, for real
When finding a relevant image, it’s personal. What’s relevant for one person or brand may not be relevant for another. Today, it’s becoming a must to personalize search results, and AI has helped evolve search to make this happen.
Applying AI can help find the best images that truly match a brand’s aesthetic or style. Machines can look at photos, extract relevant information, and become smart enough to understand what a brand wants. For example, researchers can train AI to find images within a library that match a brand aesthetic to help consultants gain access to on-brand images for the companies they work with.
The real ROI of AI
The hype of AI is real, but so is the ROI. Online ad spend has already overtaken TV spend (which itself is a $72 billion industry), with more than $40 billion spent on social platforms alone. Computer vision technology is also leveraged to deeply understand the impact of visuals on things like conversion, retention, and customer acquisition. AI is perfectly poised to deeply affect marketing, from sourcing and editing to predicting the performance of campaigns and visuals.
It might seem like these technologies replace human creators and consumers, but we believe they’re empowering us by offering creative freedom. Machines won’t take over the work of photographers, they will simply make it quicker and easier to complete. The boring stuff goes out the window, giving creators more time to focus on what matters: creation.
This piece was coauthored by Ramzi Rizk, cofounder and chief technology officer of stock photo provider EyeEm, and Appu Shaji, its head of research and development.