engineering/www.remarkbox.com
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Replace AI with machine learning in all user-facing text

Browse source
This commit is contained in:
russell@unturf.com 2026-02-02 15:17:38 -05:00
parent dd7f85b478
commit 420c0df7c1
2 changed files with 9 additions and 9 deletions
Download patch file Download diff file Expand all files Collapse all files

10
privacy-policy.html
View file

@ -106,12 +106,12 @@ Remarkbox uses automated systems to detect and manage spam content. This include
<ul>
<li><b>Content scoring:</b> comments are analyzed for spam signals such as link density, known spam patterns, duplicate content, posting velocity, and content length anomalies. A numerical spam score is stored alongside the comment.</li>
<li><b>AI relevance checking:</b> when enabled by the namespace owner, comment content may be sent to a third-party AI inference service (currently Hermes, hosted at hermes.ai.unturf.com, operated by UncloseAI) to determine whether the comment is relevant to the discussion. The AI receives the comment text, the thread context, and the namespace or page URL. The AI's assessment (relevant or irrelevant) and a brief explanation are stored alongside the comment.</li>
<li><b>Namespace owner control:</b> namespace owners may disable AI-based spam filtering for their namespace at any time via namespace settings. Basic spam scoring based on content patterns still applies when AI filtering is disabled.</li>
<li><b>Machine learning relevance checking:</b> when enabled by the namespace owner, comment content may be sent to a third-party machine learning inference service (currently Hermes, hosted at hermes.ai.unturf.com, operated by UncloseAI) to determine whether the comment is relevant to the discussion. The system receives the comment text, the thread context, and the namespace or page URL. The system's assessment (relevant or irrelevant) and a brief explanation are stored alongside the comment.</li>
<li><b>Namespace owner control:</b> namespace owners may disable machine learning-based spam filtering for their namespace at any time via namespace settings. Basic spam scoring based on content patterns still applies when machine learning filtering is disabled.</li>
</ul>
<p>
Comments flagged by the automated moderation system may be held for manual review by namespace moderators. Spam scores and AI explanations are visible to namespace moderators to assist with moderation decisions.
Comments flagged by the automated moderation system may be held for manual review by namespace moderators. Spam scores and machine learning explanations are visible to namespace moderators to assist with moderation decisions.
</p>
<h4>Marketing: Data Collection</h4>
@ -149,7 +149,7 @@ Based on the consent you've given, we may permit the following companies to use
<li><b>Remarkbox</b> - For website comments and/or "frequently asked pages"</li>
<li><b>Stripe</b> - For payment processing and subscription management</li>
<li><b>Google Analytics</b> - For website analytics</li>
<li><b>UncloseAI (Hermes)</b> - For AI-based content relevance checking, when enabled by the namespace owner. Comment text and thread context are sent for analysis. No personal identifiers (email, IP) are sent to this service.</li>
<li><b>UncloseAI (Hermes)</b> - For machine learning-based content relevance checking, when enabled by the namespace owner. Comment text and thread context are sent for analysis. No personal identifiers (email, IP) are sent to this service.</li>
</ul>
<h4>What kinds of automated decision-making or profiling do we perform with your data?</h4>
@ -220,7 +220,7 @@ The Services are operated in the United States. If you are located outside of th
</p>
<p>
When AI-based content moderation is enabled, comment text and thread context may be processed by the Hermes AI inference service, which is also operated in the United States.
When machine learning-based content moderation is enabled, comment text and thread context may be processed by the Hermes machine learning inference service, which is also operated in the United States.
</p>
<h4>Privacy Policy Updates and Revisions</h4>

8
terms-of-service.html
View file

@ -81,13 +81,13 @@ You agree not to use the Service to:
<h4>6. Automated Content Moderation</h4>
<p>
Remarkbox uses automated systems to detect and manage spam. This includes content scoring algorithms and AI-based relevance checking (powered by Hermes, a third-party AI inference service operated by UncloseAI). Comment content and thread context may be sent to the AI service for analysis when this feature is enabled by the namespace owner.
Remarkbox uses automated systems to detect and manage spam. This includes content scoring algorithms and machine learning-based relevance checking (powered by Hermes, a third-party machine learning inference service operated by UncloseAI). Comment content and thread context may be sent to the machine learning service for analysis when this feature is enabled by the namespace owner.
</p>
<p>
Content flagged by the automated system may be held for moderator review, labeled as spam, or rejected. Namespace owners and moderators can override automated decisions. Namespace owners may disable AI-based moderation for their namespace at any time.
Content flagged by the automated system may be held for moderator review, labeled as spam, or rejected. Namespace owners and moderators can override automated decisions. Namespace owners may disable machine learning-based moderation for their namespace at any time.
</p>
<p>
Remarkbox stores spam scores and AI explanations alongside comments to support moderation transparency.
Remarkbox stores spam scores and machine learning explanations alongside comments to support moderation transparency.
</p>
<h4>7. Rate Limits</h4>
@ -110,7 +110,7 @@ Remarkbox provides a JSON API for programmatic access. Use of the API is subject
<h4>10. Privacy</h4>
<p>
Your use of the Service is also governed by our <a href="/privacy-policy.html">Privacy Policy</a>, which describes how we collect, use, and share your personal data, including the use of AI-based content analysis.
Your use of the Service is also governed by our <a href="/privacy-policy.html">Privacy Policy</a>, which describes how we collect, use, and share your personal data, including the use of machine learning-based content analysis.
</p>
<h4>11. Data Export and Account Deletion</h4>