Skip to content
⌘K
InboxAsk a human

Risk catalogue

Updated by Hugo

This page lists every risk available in Unbox, grouped by category. Risks marked requires configuration need additional information from you before they can work correctly. See the note under each one for what you'll need to provide.

To learn how to enable or disable risks, or to change how they behave, see Setting up your risk configuration.

How the catalogue is organised

Risks fall into three groups.

  • Core safety – a baseline set, always on for every brand and aligned to the World Federation of Advertisers (WFA) and Global Alliance for Responsible Media (GARM) brand safety standards. These cannot be disabled.

  • Enhanced risks – the configurable set. You switch these on according to your brand's requirements, and set the severity, review mode and policy for each. Every category below other than Core safety is part of this set.

  • Custom risks – bespoke models built with you for legal, regulatory or compliance needs specific to your brand. These are requested through Unbox rather than switched on yourself.

Core safety

These risks are always enabled and cannot be disabled. They cover content that falls outside acceptable limits regardless of brand, category, or campaign context.

Risk

What it detects

Harassment

Content that bullies, intimidates, demeans or targets an individual or group.

Hate

Content that expresses, incites, or promotes hate based on race, gender, ethnicity, religion, nationality, sexual orientation, disability status, or caste.

Illicit activity

Content that includes instructions or advice that facilitate planning or executing illegal acts.

Self-harm

Content that promotes, encourages, depicts, or instructs acts of self-harm.

Sexual content

Content meant to arouse sexual excitement, promote sexual services, or depicting explicit or inappropriate sexual behaviour.

Sexual content involving minors

Sexual content featuring an individual under 18. Treated separately from general sexual content given the severity.

Violence

Content depicting death, violence, or physical injury, including graphic detail or cruelty to people or animals.

Stereotyping & ridicule

Stereotyping or ridicule based on race, gender, religion, or similar protected characteristics.

Unethical behaviour

Content promoting antisocial, dangerous, or self-destructive behaviour such as violence or drug abuse.

Controversial subjects

Inflammatory or divisive topics including fake news and political manipulation.

Claims & messaging

Risks in this group cover what creators say about your product and the language they use to say it: false, misleading, exaggerated, or prohibited claims, and words you don't want associated with your brand.

Risk

What it detects

Notes

Harmful product claims

False, misleading, or dangerous claims about the product.

Prohibited claims ⚙

Specifically prohibited claims such as disease treatment or hangover cures.

Requires configuration. You need to provide a list of the specific claim types that are prohibited for your brand or product.

Manipulative purchase tactics

False urgency, fake scarcity, fake countdowns, or fake social proof used to pressure purchase decisions.

Banned keywords ⚙

Prohibited words or phrases appearing in the video.

Requires configuration. Provide your list of banned keywords. Detection only applies to the terms you specify.

Swear words

Profanity or vulgar language used in the video.

Covers profanity generally, so you don't need to list swear words individually under Banned keywords.

Third parties

Risks covering competitor and retail partner mentions.

Risk

What it detects

Notes

Competitor disparagement ⚙

Negative references to or comparisons with competitor brands.

Requires configuration. Provide the list of competitor brands and products you want to monitor.

Retailer disparagement

Negative references to or comparisons with retail partners.

Third-party logos ⚙

Competitor or third-party brand logos appearing in the video.

Requires configuration. Provide the list of third-party brands whose logos should not be flagged.

Audience & participants

Risks relating to who appears in the video.

Risk

What it detects

Appeal to minors

Content explicitly targeting children under 13.

Minor consuming product

Content showing a child holding or consuming the product.

Minor spokesperson

Content featuring a child as the primary speaker or on-screen character.

Vulnerable groups

Content shows the product being used by children, pregnant individuals, or adults clearly older than 70.

Animals present

Content showing real animals or pets.

Product depiction

Risks relating to how the product itself is shown or used.

Risk

What it detects

Overconsumption

Content depicting excessive consumption of the product.

Food misuse

Content showing disrespectful handling or wasting of food products.

Challenge consumption

Content presenting product consumption as a challenge or game.

Alcohol & tobacco

Content mentioning or showing alcohol or tobacco products.

Authenticity & disclosure

Risks covering how content is presented and whether required disclosures are present.

Risk

What it detects

Notes

Missing disclosure ⚙

Required disclosure text overlays or hashtags are absent.

Requires configuration. Specify the disclosure text or hashtags your campaigns require (e.g. #ad, #sponsored).

AI-generated content

Content that is likely AI-generated, detected via generative-image inconsistencies and related signals.


Music

Risks covering intellectual property and licensing.

Risk

What it detects

Notes

Unlicensed music

Detects use of music outside TikTok's Commercial Music Library.

Explicit lyrics

Song lyrics referencing sexual behaviour, unethical conduct, or violence. Also catches audio that would trigger other core safety categories.


Custom risks

Anything not in this catalogue can be built as a custom risk, scoped with your Unbox team against your own brand guidelines. Common examples include:

  • Healthcare and ingredient claims specific to a regulated product range.

  • Owned and associated brand logos, so sibling brands are not flagged as third parties.

  • A brand code of conduct, where your own standards go beyond the standard catalogue.

Custom risks are requested using Request custom risk on the Risks tab, and are subject to an additional charge. See Setting up your risk configuration for how the process works.