IAB Brand Safety and Suitability Guide (2020)

Tags: Frameworks

TL;DR

IAB’s 2020 guidance separates universal brand safety (e.g., violence, hate, illegal content) from brand-specific suitability thresholds.
The AI era adds a new quality dimension: how content is made and overseen, not just what it’s about.

Why it matters for HK marketers: Suitability policies must evolve to include AI-content quality signals alongside classic category exclusions.

Core definitions (2020)

Brand safety: Avoiding universally harmful categories (violence, hate speech, illegal content).
Brand suitability: Tailored tolerance for context, tone, and adjacency based on brand values and risk appetite.

What’s changed in the AI era

New axis of risk: AI-generated content can be “safe” by category yet unsuitable due to inaccuracy, spamminess, or low value.
Quality signals required: Evaluate human oversight and editorial review of AI-assisted content, not just topics and keywords.

Policy updates to make

Include AI-content provenance and oversight in suitability taxonomies.
Require partners to disclose AI editorial processes and verification coverage.

68% of respondents want to avoid inaccurate or hallucinated AI content entirely.

63% flag spam-like or cluttered AI experiences as unacceptable adjacency.

So what for marketers

Amend your brand suitability matrices to score AI-content quality and publisher oversight, and align verification partners to enforce these thresholds.

Sources:

Article1_Brand_Safety_Suitability_Redefined_v2.docx

← Back to Knowledge Base