Predicting Your Watching Preferences: Netflix's AI Experiment
Netflix's earlier belief in relying solely on algorithms to predict user preferences has not yielded the desired results. As Reed Hastings, the founder and former CEO, envisioned in 2017, users would be presented with a minimal grid of content choices, Netflix has realized that a more sophisticated approach is needed.
AI Discover: A New Content Recommendation Feature
An experimental AI-powered feature, "AI Discover," is now being rolled out to a small group of U.S. users. This feature prompts users to select their preferred programming genre, such as "Family Movie Night," "Scare Me," or "Quick Watches," and then offers a curated cluster of video options. It appears as a navigation option in the top bar.
Beyond Algorithms: Previous Attempts at Content Discovery
Netflix has previously explored alternatives to its algorithmic rows, such as:
- Daily Top Lists: In 2020, Netflix introduced daily rankings of the most-watched movies and shows, acknowledging the popularity of discovering trending content beyond personalized recommendations.
- Categories Menu: In 2022, a dedicated "Categories" section was added, though it was later removed during an interface overhaul.
- Surprise Me Button: The "Play Something" or "Surprise Me" button aimed to combat decision paralysis by randomly selecting content from watchlists or personalized recommendations. However, it was discontinued due to low usage.
Netflix's AI Approach: A Test Kitchen for New Ideas
Netflix's AI Discover is not about conversational AI like that of Amazon or Roku. Instead, it offers tailored content batches based on user preferences. The company views this as a "test kitchen" for new discovery methods, where user feedback will determine the next steps.
Addressing Engagement Concerns
The introduction of AI Discover comes as Netflix faces reported engagement stagnation among subscribers. Recent financial forecasts, workforce cuts, and executive departures highlight the challenges the company is facing. While new navigation features might not solve all issues, they could help users navigate the vast content library more effectively.