In plain terms
- Volumes of automated and human moderation over the reporting period.
- Number of appeals received and how they were resolved.
- Methodology used and how long the underlying data is kept.
Non-binding summary, provided for readability. Only the full text below is legally binding.
Back to ZlobbyLEGAL
Transparency Report (DSA)
In accordance with Article 15 of EU Regulation 2022/2065 (the Digital Services Act, or DSA), Zlobby publishes statistics on its content moderation activity. This report also explains how our moderation system works, the role of artificial intelligence, and your rights to appeal. The mandatory consumer-protection law of your country of residence remains applicable.
Last updated: 16 June 2026
Period covered
From 9/18/2026 to 10/18/2026 (last 30 days, updated every hour).
Automated moderation (AI)
- Total scanned: 161
- Automatically approved (≥90% confidence): 127
- Automatically rejected (≥95% confidence): 25
- Confirmed by a human moderator: 3
- Overturned (AI false positive): 2
User appeals (Art. 20 DSA)
- Appeals filed: 0
- Original decision upheld: 0
- Decision overturned in favour of user: 0
How moderation works
Zlobby brings together two worlds: a social network and a Dating mode that is enabled only through explicit opt-in. All public or semi-public content can be moderated: posts and post images, comments, bios, avatars, gallery photos, private messages and lobby messages. A profile photo is mandatory at sign-up and is systematically analysed by AI before it goes live.
Automated analysis is carried out by our processor Anthropic PBC (the Claude model), which reviews the submitted text and images. Each decision follows three levels: safe, flagged, or rejected, each with a confidence score. When in doubt, content is routed to human review. Detected categories include, among others: nudity or sexual content, violence, hate speech, harassment, spam or advertising, illegal content, impersonation, content involving minors, self-harm, drugs, scams or fraud, and exposure of personal information.
Your rights (Art. 17 & 20 DSA, Art. 22 GDPR)
When a moderation decision affects you, you receive a statement of reasons (within the meaning of Article 17 of the DSA) specifying the content concerned, the reason, and whether the decision was made by AI or by a human. You may request human review of any automated decision (Article 22 of the GDPR) and contest the decision within 30 days. A report from another user triggers a fresh scan and opens an appeal flow for both the author of the content and the person who reported it.
Sanctions and refunds
An actioned report results in either content removal, a warning, or account suspension. AI rejections for non-severe categories (spam, duplicate, off-topic) do not trigger automatic warnings in order to minimise false positives. The permanent ban threshold is 10 warnings (lowered for the most serious categories). Credits spent on a post rejected by the AI are refunded automatically. In the event of a ban, active Stripe subscriptions (Pro, Legend) are cancelled automatically; see our refund policy.
Methodology and retention
The figures above are extracted live from our moderation tables. Moderation audit-trail records (decisions, appeals, reports, actions) are retained for 5 years for DSA compliance. For any question about this report or a decision affecting you, write to contact@zlobby.com. You may also lodge a complaint with the data-protection authority of your country of residence.