A surge of low-grade, AI-generated video has hit YouTube at scale, and major advertisers are now blocking their campaigns from running against it, Marketing Brew reported.

The category, widely derided as "AI slop," covers mass-produced, often faceless videos built from generative models — usually paired with clickbait thumbnails, repetitive narration, and stories that are misleading or wholly fabricated.

What counts as "AI slop"

The format took off in 2023 and accelerated through 2024 as text-to-video tools got cheaper.

Channels pushing hours of synthetic content per day have grown into YouTube's mid-tier, filling niches like:

• Children's story narration
• History explainers
• "Top 10" countdowns
• Finance and "manifestation" shorts

YouTube has not disclosed how much of its library is now AI-generated, but the volume is visible in any recommendation feed.

Why marketers are pulling spend

Programmatic ad buying runs on placement signals: keywords, channels, content categories. When those signals land on slop inventory, advertisers pick up brand-safety violations.

DoubleVerify, Integral Ad Science (IAS), and creator-verification vendors have all extended their brand-safety categories in the past year to cover "AI-generated" or "synthetic media" as a discrete flag.

Marketers can apply that flag as a hard block across their YouTube buys through demand-side platforms.

What buyers give up

The reach tradeoff is real. A blanket "AI" block removes single-digit to low-double-digit percentages of available impressions, depending on vertical.

For performance advertisers buying on cost-per-view, blocking cheap synthetic inventory raises effective CPMs. Several agency buyers have reportedly opted for narrower blocks instead:

• AI content plus low engagement signals
• AI content plus misinformation categories
• Known-channel exclusions over category-wide bans

Where disclosure falls short

Google started requiring AI-content disclosure on YouTube in late 2024, rolling out a label letting creators tag videos as "made with AI" or "altered or synthetic."

The label covers realistic-looking faces, voices, and scenes. The system relies on self-attestation, and enforcement has been thin.

Creators have bypassed the rule by using AI for non-realistic material — animation, slideshows, simple commentary — that doesn't trigger disclosure.

Automated detection of generative content remains an unsolved problem across the industry; detection tools struggle with films and shorts under three minutes.

The knock-on for creators

Channels that built audiences on AI-assisted scripting or b-roll are weighing a return to on-camera presentation. Larger creators report better CPMs once they publish under a real face and voice. Smaller channels lose scale.

Some have rotated to TikTok and YouTube Shorts, where face presence is less central to the format.

The next test will come when YouTube begins auto-detecting AI content rather than relying on self-labeling. Google has not announced a date.