IAS Low-Quality GenAI Avoidance
Tags: Tools & Platforms, Statistics
TL;DR
- IAS classifies and enables avoidance of low‑quality AI‑generated content ("AI slop") across digital inventory.
- Large‑scale testing shows higher success rates and lower costs when bids avoid AI‑slop supply and prioritize higher‑quality inventory.
Why it matters for HK marketers: Turning on this control and feeding its signals into bid models can improve performance and reduce waste in Hong Kong media buys.
What it is
- Product capability: IAS identifies low‑quality GenAI content and allows buyers to avoid it pre-bid or steer optimization away from it.
- General availability: Marketed as generally available in May 2026 ("Cut the AI Slop" release).
Performance evidence (aggregate)
- Dataset: 1B+ impressions over four days.
- Result: Non‑AI‑slop inventory delivered a higher success rate and lower cost per success versus AI‑slop inventory—improving outcomes without paying a premium.
How buyers use it
- Pre-bid activation: Exclude AI‑slop segments to minimize waste.
- Optimization signals: Pass IAS quality signals into DSP algorithms so models learn toward quality cohorts that convert more efficiently.
- Measurement: Track success rate and cost per success pre/post activation to verify incremental lift.
+49% success rate on inventory not classified as low‑quality AI-generated content (IAS, 1B+ impressions).
-24% cost per success on the same higher‑quality inventory vs. AI‑slop inventory (IAS).
So what for marketers
Ensure IAS Low‑Quality GenAI Avoidance is enabled pre‑bid and its signals are integrated into optimization logic; validate gains via pre/post tests on success rate and cost per success.
Sources:- Article2_Detection_at_Scale_v2.docx
← Back to Knowledge Base