This article explains how the KICK moderation team assesses content, decides on enforcement and handles appeals.
Our Community Guidelines set the rules; this explains how we apply them.
1. How We Moderate
Moderators review live streams, VODs, clips, chat and channel content flagged by user reports, automated detection, proactive monitoring, and notices from regulators, trusted flaggers and law enforcement. Most of it is live and unscripted, so two
moments that look the same can mean very different things.
Identify: Content is flagged by a report, automated detection, monitoring, or a notice from a regulator, trusted flagger or authority.
Assess: A moderator reviews the content in context and rates intent and outcome to reach a risk rating.
Enforce: An action proportionate to the rating and the standards for that Community Guideline.
Escalate: The most serious cases go to Trust & Safety and Legal and, where the law requires, to authorities.
Review: The creator is notified and can appeal.
Moderation is context-aware rather than a fixed list of banned words or images. Automated tools help find content; trained moderators review it and decide on enforcement.
2. The Assessment: Context, Intent, Outcome
Every assessment answers three questions in order: what was happening, what did the streamer do, and what was the effect?
2.1 Context
Category: Does the stream category match what is shown?
Labelling: Is the channel labelled 18+ where content is for mature audiences?
Timeline: What happened before and after, not only a single frame in isolation.
Predictability: Could an outcome reasonably have been predicted.
Location: Local laws and standards apply
History: Previous warnings and enforcement for similar conduct.
2.2 Intent And Outcome
Repeat conduct moves up the intent scale: the same behaviour after a warning is at least reckless. Outcome is judged on harm, not report volume, and limited harm repeated after warnings can reach moderate.
3. The Risk Matrix
Intent and outcome meet on the matrix. Each cell gives a risk rating, which sets the range of enforcement.
3.1 How To Read It
A starting point: Each Community Guideline has its own standards for common scenarios, and those take priority. Child safety is stricter than the matrix.
Responsible streamers: Moderation enforcement can involve any one or combination of channel enforcement options.
Zero tolerance: Child sexual abuse material, terrorism and violent extremism, non-consensual intimate imagery and self-harm, doxxing and swatting are generally the highest priority issues where there can be an imminent threat.
Local law: Where law is stricter, enforcement may be stronger in that market, or content restricted in one country rather than globally.
Aggravating factors: Attempting to gain from the negative conduct, involving minors or people who didn't consent, or continuing after a warning notice.
Mitigating factors: Correcting the issue quickly, a first instance, and a clean channel history.
4. Enforcement Toolkit
Moderators use the smallest intervention that stops the harm and prevents it recurring, often in combination. Enforcement builds: repeat conduct starts higher.
5. Worked Examples
The same situation can land anywhere on the scale. Each example shows how the rating climbs as intent and outcome change.
5.1 A Real Fight On An IRL Stream | VIOLENCE
5.2 Exposure On A Mature Stream | SEXUAL CONDUCT
5.3 Gambling Streams | GAMBLING
5.4 Randomised Video Chat And Minors| PROTECTION OF MINORS
6. Escalation, Decisions And Appeals
Beyond the platform: Child sexual abuse material, terrorism and violent extremism, and imminent risk to life are escalated straight away to Trust & Safety and, where the law requires, reported to authorities such as child protection hotlines or law enforcement, with evidence preserved securely. Trusted flagger reports and regulator requests are prioritised.
Decisions: We tell the creator which Community Guideline a decision relates to. Creators in the EU receive a statement of reasons covering the decision, the facts and how to challenge it, and may be able to refer it to a certified out-of-court dispute settlement body.
What we keep internal: We publish this framework so creators know what to expect, but not every internal threshold, detection signal or workflow, because that detail could make enforcement easier to evade.