Position paper on artificial intelligence in the VOD sector

Ahead of the European Commission’s upcoming proposal on artificial intelligence, the European VOD Coalition outlines how AI systems are used in the VOD sector operate and how these differ from their application in other business areas.

Position paper on artificial intelligence in the VOD sector
Photo by Przemyslaw Marczynski / Unsplash

The VOD sector

The video-on-demand (VOD) sector is distinguished from other businesses such as online video-sharing platforms (VSPs) by operating within a closely curated environment with strict editorial controls. For this reason, there is no user-generated content on these services. VODs have a singular goal: help viewers find the films and TV series they are likely to enjoy. This focus on viewer enjoyment has led to an explosion of European content creation and more diverse choices in what to watch than ever before.

Historically, television has been bound by a linear program schedule: the content available is limited to what is being broadcasted at that moment. But VOD business models depend on consumers consistently finding amazing content to watch whenever the consumer wants. This gives streaming service customers the opportunity to discover new content and enables streaming services to make deeper investments in new and innovative programming.

VODs are solely focused on finding content that appeals to the individual tastes of each of their customers. Because of this, VODs have enabled European and global titles to reach new audiences across Europe and around the world, reinforcing the adage that what travels best is a good story.

Personalisation without intrusiveness

A VOD’s ability to match content with new audiences is dependent on systems that help consumers find content they may enjoy without having to manually parse through a large catalogue. These systems often use a combination of personalized recommendations, objective ranking, search, and genre-based filtering to help viewers find shows and movies they are likely to enjoy. This allows members to browse large catalogues of films and series easily and quickly such that they can make an informed decision about what to watch.[1]

Every VOD business uses a unique system to recommend content in a way that fits their individual service and catalogue. Oftentimes, in order to help viewers find something to watch that they enjoy, recommendation systems may use interactions with a service (including viewing history or how a viewer rates titles), what other viewers with similar tastes and preferences are watching, and information about the titles (such as their genre, actors, release year, etc.). VODs often include a number of manual filtering mechanisms and informational tools to help viewers make informed decisions about what to watch. They can include age filtering, parental control tools, age ratings, content descriptors, and summaries.

A VOD's only objective is to present consumers with content they are likely to enjoy. There is no inherent benefit from highlighting one genre or type of content over another. Individuals tend to appreciate a diverse variety of content and don’t tend to watch the same things all the time, VODs are incentivised to ensure there is always a diverse catalogue of content being recommended.

Recommendations also allow European content to find its ideal audience in the market where it was made, in other European countries, and around the world. This is because recommendations take a multifaceted approach to recommending content. For example, a consumer from Romania who likes romantic movies would be served well by being recommended a Romanian romantic movie. And a Romanian romantic movie would be well placed to be discovered by viewers when placed next to romantic movies from other European countries and romantic movies from around the world. Further, European content is well positioned to benefit from VOD recommendations in Europe because content producers working in the same linguistic, cultural, and artistic context as potential consumers have an inherent advantage in producing works that attract those viewers.


Content curation in the cultural and creative sector is vastly different from other business segments. An editorial approach reduces the risk of misinformation appearing on a VOD service and viewer’s variety in taste incentivizes recommendation systems to help promote exposure to broad and varied catalogues of content, including content across multiple genres, from different countries, in different languages, and with varying themes.

These incentives and attributes mitigate the risk that VODs will have a detrimental impact on European citizens. It is therefore imperative that any rules on curation and recommendation systems granularly distinguish between the environment they are operating in and for what purpose they are used. For example, enforcing an opt-out from content curation in these systems would lead to an untenable user experience that would hamper the ability to discover new content.