YouTube is cracking down on low-quality AI-generated videos, which now comprise over 20% of content shown to new users on the platform.
Rather than removing individual videos, the company is targeting accounts posting AI slop at scale to harvest ad revenue.
YouTube is also rolling out tools to detect and label deepfakes and AI-generated content, and has created a likeness protection program allowing public figures to request removal of videos using their likeness — an effort to balance moderation against accusations of political bias that plagued the platform during the COVID-19 pandemic.
What happened
YouTube has tightened rules on monetizing "inauthentic content" — including repetitive videos, template-based AI content, and AI personas discussing sensitive topics — and is focusing enforcement on accounts posting AI slop at scale rather than individual videos. According to video editing company Kapwing, more than 20% of videos YouTube's algorithm shows to new users are low-quality AI-generated content.
Why it matters
YouTube CEO Neal Mohan called combating AI slop a "top priority" in a public letter earlier this year, signaling that the platform sees low-quality AI video as a genuine threat to its brand. The company faces a familiar tension: during COVID-19, YouTube faced criticism both for allowing health misinformation and for removing it — accused of stifling free speech or playing politics. Balancing moderation of AI content, especially political deepfakes, without appearing to censor one ideology over another is a test of the platform's neutrality claims.
What to watch
YouTube has launched a likeness protection program that automatically detects videos featuring a public figure's face and lets them petition for removal. The company is also automatically applying labels to realistic AI content for transparency. Political content remains a small slice of AI slop today, but YouTube's approach to deepfakes and sensitive-topic AI personas will shape how platforms manage this risk going forward.
YouTube, the dominant online video platform, is racing to manage a growing problem: AI-generated videos that are cheap to produce, often low quality, and designed to trick viewers and harvest ad revenue. According to video editing company Kapwing, more than 20% of the videos YouTube's algorithm shows to new users fall into this category of low-quality AI-generated content.
CEO Neal Mohan declared combating "AI slop" a "top priority" in a public letter earlier this year. The company has responded by tightening its monetization rules around what it calls "inauthentic content" — including repetitive or template-based videos, AI-generated videos with no clear narrative, and content where AI personas discuss sensitive topics. Rather than remove individual videos, YouTube has focused enforcement on accounts posting such content at scale, attempting to cut off the financial incentive driving the spam. "To reduce the spread of low quality AI content, we're actively building on our established systems that have been very successful in combatting spam and clickbait, and reducing the spread of low quality, repetitive content," Mohan wrote.
Political content represents a small slice of the AI slop problem today, but it poses an outsized reputational risk. YouTube learned this lesson during COVID-19, when the platform faced criticism both for hosting health misinformation and for removing it — accused of stifling free speech or playing politics. To navigate that minefield, YouTube has taken a more careful approach to AI-generated political content, emphasizing that it applies community guidelines consistently regardless of political ideology. The company's main focus has been targeting deepfakes, which violate its privacy policy. YouTube has launched a likeness protection program that automatically detects videos featuring a public figure's likeness and allows that person to petition for removal; the platform is encouraging public officials, celebrities, journalists, and creators to enroll. The company is also automatically applying labels to realistic AI content for transparency. "As this technology advances, we're building on those systems with industry-first tools to protect creators' likenesses, and automatically applying labels to realistic AI content for transparency," a YouTube spokesperson told Semafor.
YouTube's challenge with AI-generated content sits at the intersection of two platform pressures: maintaining brand quality and avoiding accusations of censorship. CEO Neal Mohan's public letter earlier this year naming AI slop a "top priority" reflects the scale of the problem — over 20% of new-user feeds now contain low-quality AI video. The company's strategy has shifted from policing individual videos to targeting accounts that post AI content at scale, a more efficient approach to cutting off the financial incentive (cheap ad revenue) that drives such spam.
The political dimension adds complexity. During COVID-19, YouTube faced contradictory criticism: it was faulted both for hosting health misinformation and for removing it. Any aggressive move against AI-generated political content risks renewing claims that the platform favors one ideology over another. YouTube's stated response — applying community guidelines "consistently, regardless of political ideology" and focusing on deepfakes that violate its privacy policy — is an attempt to frame moderation as technical and neutral rather than editorial. The likeness protection program and automatic labeling of realistic AI content are tools designed to do this work at scale, without human judgment calls that could appear partisan.
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytime
Ask AI anything about this article. Q&As are published on this page for other readers too.
IAPS fellow Severin Field interviewed 25 researchers from OpenAI, Anthropic, Google Deepmind, Meta, and US uni…
Anthropic announced it will add invisible, machine-readable watermarks to all content processed by Claude—not…
German enterprises are changing how they approach Microsoft technologies as the company integrates workplace…
MIT Technology Review interviewed kids aged 10 to 18 about their relationship with AI
Meta announced a $1B "Future Is For Everyone Fund" targeting communities affected by its AI infrastructure pro…
Zoox, Amazon's autonomous vehicle subsidiary, began charging passengers for driverless rides in Las Vegas on M…
The AI news that matters, in one minute each morning.
Sign up free