New Product Feature: EU Infrastructure Region
Spectrum recently expanded its global capabilities by adding a new infrastructure region in Frankfurt, Germany. We now have 4 regions around the world, to enable community platforms to detect harmful behaviors and trigger actions without latency, wherever customers are located. Our infrastructure regions include Oregon and Virginia for the US; Seoul, Korea for Asia-Pacific; and now Frankfurt, Germany for Europe-Middle East-Africa.
Our team is expert in building massively scalable big-data processing applications. Because of this, our content moderation application is extremely fast, with an internal response time in the 10s of milliseconds.
Even with such a rapid internal response time, latency introduced by going through the network can limit customers’ ability to use certain content moderation features. Adding more data centers like this one increases our geographic coverage, enabling us to better support European and global customers.
Our robust global infrastructure automatically finds the location closest to each customer, and creates backup capabilities across regions. The end result is greater scalability for our customers as they add new users across the globe.
Why It’s Exciting: Real-Time Redaction
Our enterprise-grade technology is super-fast, but network calls still take time. When a community platform requests that we analyze its user-generated content, the request travels to our API to get the analysis result, then travels back to the community platform to take action.
Placing infrastructure regions close to our customers reduces the time required to make these round-trip network calls. With lightning-fast response times, we can help customers proactively stop harmful content from being published.
In an online community, two users are having a conversation, and one user types an insult. For example, this could be one child calling another a “moron.”
Here’s how real-time redaction works:
- Detection: Spectrum’s AI behavior detection model flags the insult and returns the determination to the community platform.
- Redaction: The community platform automatically stops the insult before it can be published.
- Warning: The platform also warns the user that such insults will be blocked in the future, and that multiple infractions will trigger further actions such as suspending or blocking the user.
The ability to perform real-time redaction is valuable for communities to not just catch harmful behaviors after the fact, but proactively stop them before they are published and harm is done. Proactive redaction requires a round-trip time of <50 milliseconds. If network latency is too high and response times are too slow, real-time redaction isn’t possible. In this example, one child would still be able to insult another. Our powerful global infrastructure means our customers can rely on real-time redaction to protect their users.
Results: Protection for Your Global Community
Everyone benefits from fast, automated content moderation:
- The insult is blocked and the user who typed the insult is warned.
- The other user is protected from the insult.
- The community moderator has increased efficiency due to automated action.
- The community platform has prevented a negative interaction that could reduce user engagement and loyalty, and drive the user away.
Spectrum also offers a patented AI approach to multi-language content moderation. Detect the languages being used on your platform, and add moderation capabilities in new languages quickly and easily. Ensure coverage and gain insights across all languages and regions in your growing global community.
Contact us to learn more about protecting your global audience with content moderation:
Whitepaper | Moderating Content in Multiple Languages
Community guidelines must be enforced consistently, regardless of the language used. But updating profanity and RegEx filters is time-consuming and ineffective at keeping up with changing trends. Learn how to achieve accurate, real-time content moderation in multiple languages.
Solution Guide | Multi-Language Moderation
Our patented, real-time, automated AI solution helps you scale globally faster, at lower cost, with better results. Discover how AI content moderation can:
- Quickly scale support for new languages
- Continue learning to improve results
- Eliminate keyword filters, cost and complexity