TikTok's recommendation algorithm serves AI-generated junk content to new users at roughly three times the rate YouTube does, according to a report from Startup Fortune — a gap that should be on the radar of every advertiser buying reach on the platform.

The number matters because new-user feeds are where acquisition campaigns run. If a first-time viewer's For You page is dominated by low-quality synthetic video, the brand message sitting next to it inherits that environment. The report's framing is blunt: advertisers should be paying attention.

The term "AI slop" has moved from internet slang into trade vocabulary over the past two years. It describes mass-produced synthetic content — generated video, AI voiceovers, stitched stock imagery — built to farm engagement rather than inform or entertain anyone. It is cheap to make, fast to publish, and platforms pay for it in trust.

Why does the new-user gap matter for ad spend?

New accounts are a special case. A user who just installed the app has no watch history, no follows, no engagement signals. The algorithm has to guess, and its guesses are shaped by whatever content performs broadly in those first sessions.

That is exactly the inventory an advertiser often buys against: broad, untargeted, intent-poor audiences acquired at low CPMs. If TikTok's cold-start pipeline surfaces synthetic filler at three times YouTube's rate, the effective quality of adjacent placements diverges too — and the cheaper platform stops looking cheap.

The 3x figure also suggests a structural difference, not a random one. YouTube's recommendation system leans on longer watch sessions and channel history, which tends to reward produced content. TikTok's feed optimizes for immediate hook-and-scroll retention, which is precisely what slop is engineered to exploit.

What can advertisers actually do?

No platform-side fix ships with this data, but buyers are not without levers:

• Audit placements. Pull post-campaign reports and check what content categories ran adjacent to delivered impressions on cold-audience campaigns.
• Reprice broad reach. If feed quality on new-user inventory is materially worse, the effective CPM premium on cleaner segments deserves a recalculation.
• Watch YouTube's pitch. A 3x gap in TikTok's disfavor is a ready-made talking point for YouTube's sales team in every upcoming QBR.

How did we get here?

Generative video tools dropped the cost of producing passable content close to zero. Distributors noticed. Engagement-optimized feeds did the rest: they reward whatever holds attention for the next three seconds, and synthetic filler holds attention well because it is tuned, iterated, and A/B-tested at a volume no human editor can match.

New users are the most exposed audience because the algorithm's quality signals for them are weakest. A viewer with no history gets fed whatever the cold-start model promotes. That is the population where the 3x disparity shows up.

What happens next?

Advertisers reading this report have two reasonable responses. One: demand feed-quality transparency from TikTok at the category level, the same way they demanded brand-safety reporting on user-generated content a decade ago. Two: benchmark. The YouTube comparison gives media buyers a concrete, defensible number to bring into planning conversations.

The uncomfortable part is that engagement metrics will not flag this problem. Slop performs. Views, watch time, completion rates — all can look healthy while the content environment around an ad degrades. Only content-level audits catch it.

For now, the data point stands on its own: three times the rate, measured against the industry's default video benchmark, on the exact audience segment acquisition budgets chase hardest. Advertisers should be paying attention. That was the report's conclusion, and it is hard to argue with the arithmetic.