Case Study: Pantheon improves image moderation with Google Vision AI and Kalamuna

Pantheon flags inappropriate uploads for 6 million users with Kalamuna

Pantheon, a web platform provider, approached Kalamuna with a challenge. They needed to build a demonstration of Drupal 8 integrated with Google Cloud’s Vision API to show the value of moving to the Google Cloud Platform. The goal was to solve a real-world problem for their client, Patch.com, which faced the risk of publishing inappropriate user-uploaded images due to the impracticality of scaling human editorial review for such a high volume of content.

Kalamuna developed a custom Drupal 8 module that leveraged the Google Cloud Vision API to intelligently analyze uploaded images. The solution automatically rejected images flagged as "adult" or "violent" and sent those flagged as "racy" into a moderation workflow for human review. While the AI technology was not 100% reliable, Kalamuna successfully delivered a functional proof of concept that demonstrated the powerful potential of integrating artificial intelligence into content management systems to automate curation and scale limited human resources.

Pantheon