DOOH Content Moderation: What AI checks and what it doesn't

TL;DR

Our DOOH content moderation pairs Vision AI with human review. AI handles standard cases — explicit content, hate symbols, clear brand-IP violations. What it doesn't decide: context, irony, political gray zones, health claims without proof. Those 6% go to the human team with 24h SLA.

What the AI auto-rejects

Clear-cut cases where Vision API stops without asking: nudity above threshold, weapons in aggressive context, hate-speech symbols (predefined list), drug paraphernalia, gambling visuals without license tag, third-party trademarked logos (Disney, sports clubs, fashion brands), phone numbers in non-approved verticals.

This list is hard-coded and updated weekly. If your industry legitimately uses one of these (licensed gambling operator), there's a pre-approval path with license upload.

What the AI auto-approves

94% of creatives are unambiguous: café ad with latte art, coworking space with logo, event promo with date and venue, mobility service with car, fashion brand with model. AI matches category + visual heuristics + standard specs. Auto-approve in under 30 seconds, booking can start immediately.

What always goes to humans

Politics (any form — including positive climate messaging), health/wellness claims with effect promises, finance with yield figures, pharma of any kind, ambiguous brand references (parody, irony, homage), creatives with unclear context (e.g., a QR code with no other information).

These hit a queue our team works through hourly — weekdays within 4h, weekends max 24h. Every rejection comes with concrete reasoning and a fix suggestion, not "policy violation".

What the AI doesn't check at all

Honest: technical specs (resolution, codec, format) are a separate pipeline step. Brand consistency with your guidelines — the AI can't do that, that's on you. Data freshness (event date, sale period) — wrong dates are your risk, not a rejection reason.

Moderation cheat sheet

94% auto-approve in <30 seconds
6% manual review, <24h SLA
Pre-approval available for regulated industries
Concrete reasoning on rejection
Re-submit after fix without re-booking
Escalation path to moderation lead

DOOH Content Moderation: What the AI checks and what it doesn't

Related Articles

Brand Safety in DOOH: How adyoutiser keeps your ads away from bad neighbors

94% auto-approve at adyoutiser: How our moderation actually works

What happens to a rejected DOOH creative? The review path at adyoutiser