Netflix unveils AI Discover, a feature that allows its subscribers to describe in plain language what they feel like watching. Still experimental, it is starting to be rolled out gradually to US households, with no known date for a wider release yet.
A natural language search
Requests such as ‘Scare me’, ‘steamy series’ or ‘films about purchasing power’ can now be made directly on the platform. The goal is to match these cravings with programs in the catalogue, whereas traditional search only responds to keywords. A search for ‘horror series’ yields results, but without taking the user’s viewing history into account, which can lead to suggesting a series already seen or a children’s program one does not wish to watch again.
AI Discover promises more targeted suggestions based on the precision of the request. The more detailed the query, the more detailed the response. For instance, ‘series about English aristocrats who love drugs’ could lead to the series The Gentlemen. The feature appears on the homepage, next to tabs like ‘My Netflix’.
For now, it offers a selection of thematic suggestions, slightly more precise than the homepage ranking, which is itself customizable. The idea is to give the subscriber a starting point, before the system learns their preferences over time.
A feature set to evolve
In remarks reported by The Hollywood Reporter, a Netflix spokesperson describes a technology based on ‘a new layer of artificial intelligence’ developed by the company. According to them, it ‘reacts in real time to feedback from our subscribers, while continuously learning’, in order to ‘understand their current cravings’ and offer ‘new avenues for discovery, beyond predefined content’. They add: ‘Tell us what you’re looking for and the experience adapts.’
The subscription video-on-demand service specifies, however, that the feature can still change over time, and that it is an evolving process. It could therefore take another form in a few months. Netflix also assures that AI Discover will not be used to influence future productions based on the general cravings conveyed to the AI.
This initiative comes in a particular context for the platform, which has notably led to several layoffs. The group must retain its subscribers, as its number of users increases without it always managing to keep them long-term, particularly from one season to the next.
