As a leader in the content moderation industry, WebPurify always strives to offer clients a significant range of cutting-edge moderation products and services. Known for our popular image moderation and video moderation services, which are performed by real-live, highly trained professional humans, and leveraging over 10 years of experience and data, we are thrilled to add our new AIM service. Combining the latest in machine learning and computer vision technologies, AIM is a simple-to-use and accurate automated intelligent moderation service for images that focuses on identifying offensive content. With this addition, WebPurify also provides an exciting and unique hybrid human and automated moderation service option for clients.
How AIM Works
The AIM solution expands WebPurify’s suite of services to provide a budget-friendly choice for clients who have a large quantity of images and require real-time automated moderation. As images are submitted, AIM responds with a percentage probability of categories such as nudity, hate symbols, offensive gestures, weapons, drugs, and more in each image. AIM is incredibly fast and extremely low cost at just $0.0015 per image.
Combined AIM and Human Service
This new automated moderation service can be seamlessly integrated with WebPurify’s live moderators. This means that images with a high probability of nudity or other offenses are automatically identified before reaching WebPurify’s human team. Clients can opt to buy moderation credits and split them between the automated and human services however they may like based on their budgets and image volumes. See how one client utilized our hybrid approach below.
Client with a problem: A high-volume dating site was utilizing a post-moderation process (moderation occurs after images go live) since their users wanted images posted immediately. The site owners were allowing unscreened images to go live on their site and were constantly behind in their efforts to take down offensive content as soon as possible. With volumes of 150,000+ user-generated images per day, they were wasting the company’s resources and team’s precious attention that could be better focused elsewhere.
AIM-human hybrid solution to the rescue: By using WebPurify’s AIM-human hybrid service, they were able to move to a pre-moderation solution where all image submissions are immediately checked for nudity, hate symbols, and weapons. Any images identified by AIM with a very high probability of containing these offenses are automatically rejected. Any images with a very low likelihood of these categories are permitted on the site, but queued up for WebPurify’s live team to review within five minutes in case other violations exist, such as violence. All images that fall in between are not permitted on the site until reviewed by WebPurify’s live team within five minutes as well. Even these images are in order of risk probability, so the more threatening images go to the back of the queue, allowing the likely harmless images to be reviewed and posted sooner.
The result: A significant percentage of the offensive images are immediately rejected and the volume of images requiring human review is significantly reduced, saving time, money and effort. Only a small percentage of images are held back for live moderation, but are still moderated rapidly and in order of priority.
A One-of-a-Kind Moderation Service
None of the automated moderation services that are available on the market provide a hybrid solution like WebPurify’s that combines a highly trained in-house human team with artificial intelligence. The same can be said of the human moderation services out there, most of which crowdsource and do not use their own in-house, trained personnel to review images like we do.
While there will always be a need for detailed human image moderation, this new, integrated technology makes the process more efficient and reduces the image volume that requires human review.
Ready to get started? You can find more information on WebPurify’s image moderation services here.