Jigsaw and Tech Against Terrorism developed Altitude in collaboration with the Global Internet Forum to Counter Terrorism (GIFCT).
Altitude is now fully supported and maintained by Tech Against Terrorism. Tech Against Terrorism will continue to make technical improvements to make Altitude more effective and easy to use as threats evolve. Altitude users will get all of those enhancements and context automatically.
Altitude is free for platforms to use. It is also offered as part of Tech Against Terrorism’s Bronze tier partnership. Learn more and request access to get started today.
All content uploaded to the tool remains archived locally, even when you remove it from your platform. This allows you to review this content later and retain a copy for archiving. If you want to delete it from Altitude, you will need to take further steps to remove it from the database.
Our tool prioritizes all cases for review to help you focus on the most critical content first. This is done based entirely on flagger metadata. Here's how it works:
Each case is supported by 1+ signals. Each case receives a score from 0-10, calculated by adding its Confidence and Severity scores, which are derived from the supporting flagger signals:
- Confidence: How likely the content is to be a true threat, based on metadata like whether the signal was provided by a human flagger or whether the flagger is considered trusted.
- Severity: How urgently action is needed, based on metadata like whether the flagger marked the signal as part of an imminent credible threat.
Example:
| Case | Confidence | Severity | Total Score | Priority level |
|---|---|---|---|---|
| Terrorist attack threat from trusted source | 3 | 7 | 10 | High |
| Violent image | 1 | 3 | 4 | Medium |
| User report without metadata | 1 | 1 | 2 | Low |
Note
The severity score is heavily weighted to ensure that cases containing critical threats are always addressed first.
Cases are ranked from highest to lowest score. In the frontend they will be labeled by their flagger priority level ("High priority", "Medium priority", or "Low priority"). If we there is no flagger data available for scoring, the case will be labeled "N/A".
Sharing Diagnostics is a feature within Altitude that creates a feedback loop between online platforms and sources of content alerts (e.g. counter-terrorism databases like Tech Against Terrorism’s Terrorist Content Analytics Platform (TCAP) and GIFCT's Hash-Sharing Database). When users make decisions on flagged content, those decisions are securely recorded and shared with the alert sources via APIs. This feedback helps the alert sources refine their content analysis, leading to improved accuracy and effectiveness in flagging harmful content.
The process is passive and privacy-focused. Decisions are shared in the background weekly (configured in signal-service/src/taskqueue/config.py) without requiring user action, and platform owners can control which alert sources receive their data. Additionally, only relevant decision data is shared, and any unnecessary information is removed to protect user and company privacy. Specifically, only the identifier of the content (e.g. URL) and the decision taken (e.g. BLOCK, APPROVE) is included.
While primarily focused on improving the accuracy of content alerts, Sharing Diagnostics also promotes transparency by allowing alert sources to demonstrate their effectiveness to stakeholders. It fosters collaboration between platforms and alert sources, strengthening the collective effort against online extremism.
Altitude does not provide authentication, as most platforms already have a preferred method of authentication, like Single Sign-On. You'll need to take care of authentication yourself. The Altitude gateway (see gateway/) uses nginx, to which you could add basic or external authentication.
Yes. Altitude does not provide user accounts, but multiple users can use the tool at the same time.
We understand that not all platforms deal with content the same way. Maybe you block content or take another action instead of removing it, and want the action buttons in the UI to reflect that. There is currently no way to configure this type of language directly in the UI. However, you can update it in the underlying config at client/src/app/admin/config.ts.
Altitude can be extended to integrate with more sources. If you'd like to add a new integration, we welcome contributions!