A video streaming service wants to suggest new shows to each subscriber based on patterns across what similar subscribers have watched and rated, aiming to increase watch time without asking subscribers to fill out a preference survey. Which type of AI solution is being described?

The correct answer is the recommendation system that suggests shows based on similar subscribers' behavior. The recommendation system analyzes patterns in what similar users have watched, rated, or interacted with and uses those patterns to suggest new items to an individual user, which matches the streaming service's goal of surfacing shows a subscriber is likely to enjoy based on similar subscribers' behavior, without requiring an explicit survey. The anomaly detection model that flags unusual viewing behavior is incorrect because anomaly detection is built to flag data points that deviate sharply from an expected pattern, such as unusual account activity, and has no mechanism for generating personalized content suggestions from shared viewing patterns. The time-series forecasting model that predicts next month's total subscriber count is incorrect because a time-series forecasting model predicts a future numeric value, such as total subscriber count next month, which is a company-wide business metric and not a per-subscriber recommendation of which shows to watch next. The image classification model that categorizes each show's cover art is incorrect because image classification assigns a category label to an image, such as classifying cover art by genre, which could support a recommendation system as one input but on its own does not use viewing and rating patterns across subscribers to generate personalized suggestions. Because the described behavior is generating individualized suggestions from patterns in similar users' activity, a recommendation system is the correct classification of the solution.