How to Use YouTube’s New AI Update to Take Total Control of Your Video Recommendations

Why It Matters

By giving viewers direct control over recommendations, YouTube aims to boost engagement and ad relevance, while the AI labeling initiative addresses growing concerns over synthetic media transparency.

Key Takeaways

• Custom feeds let users prompt AI for personalized video streams
• Feature rolls out to all US mobile users in English
• Requires enabled watch history, keeping Google tracking user activity
• AI-generated labels will flag photorealistic, AI‑altered videos
• Prompt editing refreshes feed instantly, enhancing content discovery

Pulse Analysis

YouTube’s new custom feeds harness Gemini’s generative capabilities to transform how users discover content. Rather than relying solely on the platform’s opaque recommendation engine, viewers type a simple prompt—anything from "relaxing meditation" to "off‑beat indie music"—and receive a continuously refreshed feed tailored to that request. The feature is currently limited to English‑language videos on the mobile app for U.S. accounts, and it requires watch history to stay active, ensuring the AI can draw on past behavior while still honoring the explicit user direction.

From a business perspective, the move could revitalize watch time and ad impressions by aligning video streams more closely with moment‑by‑moment user intent. Advertisers stand to benefit from higher relevance scores as viewers engage with content they explicitly asked for, potentially driving up CPMs. Moreover, the hybrid model—combining algorithmic signals with user‑generated prompts—positions YouTube against competitors like TikTok and Instagram Reels, which already emphasize creator‑driven discovery. Early adoption metrics suggest a modest lift in session length for participants, hinting at broader revenue upside if the feature scales globally.

However, the rollout also raises privacy and regulatory questions. Because custom feeds depend on enabled watch and search histories, Google continues to collect granular user data, a point of scrutiny for privacy advocates. Simultaneously, YouTube’s decision to auto‑label AI‑altered videos reflects mounting pressure to combat deep‑fake misinformation. Together, these steps illustrate a balancing act: enhancing user agency and platform trust while maintaining the data foundations that power its advertising engine. As the feature expands, its impact on engagement, ad strategy, and compliance will be closely watched by the digital media industry.