MediaWave Actionable Insights and Industry News for Media Professionals
Broadcast TV faces digital challenges, but it can learn from digital, too.

How Artificial Intelligence Is Helping Make TV A Better Experience

January 22nd, 2019   ||    by Alan Wolk

Artificial Intelligence (AI) in all its forms, was front and center at CES this year. Given that it is indeed the focus of future consumer electronics, its prominence was quite appropriate.

AI is currently powering everything from smart speakers like Alexa to self-driving cars, and as the technology itself improves and learns from its mistakes, we will likely see more and more uses and applications, possibly even more streamlined versions of this year’s CES standout, a machine that automatically folds laundry.

The television industry has also discovered AI and there are already numerous use cases.

The most popular one thus far is the use of voice to enable tune-in and discovery. Users can ask a smart speaker, remote or even a smart TV to tune in to a specific show for them or to search for a specific type of program.

Devices are even learning how to make recommendations, learning their owners’ preferences and viewing habits so that they can showcase the programs they want to watch. (To wit: Hulu Live TV, which lets everyone in the family have their own account, knows I watch a lot of Brooklyn Nets games, and so when a game is on, it is usually the first thing that pops up on my home screen.)

AI can also be used to help create playlists that keep viewers engaged. Start-up IRIS.TV uses an AI-based engine that works with IBM’s Watson to help it understand a viewer’s preferences and choose the videos they might want to watch next. So that if someone is watching videos around the Olympics, for instance, and the system knows they are in Detroit, it might serve up a video about a hockey player from Detroit, even if the user hasn’t yet watched any hockey videos.

The same type of system can be used to serve up advertising too, helping networks to understand viewer preferences and what sorts of products they may be interested in. That allows for greater automation, since the AI system can also be programmed to check for brand safety and other issues before surfacing an ad. It’s the sort of system that can be used for a number of purposes, from better targeting to frequency capping to promotion.

Getting It Right vs Getting It Wrong

The problem with AI for recommendations is that it needs to be right far more often than it is wrong, or else users will give up on it. Since what to watch is such a subjective decision, recommendations may not be the best use case for AI, at least not at first.

Advertising seems a much better use of the technology. AI can determine which products correlate with a viewer’s preferences, which ads they’ve watched all the way through and which ones they’ve changed the channel in the middle of. It can quickly figure out if they’ve seen too many car commercials in the past few days or determine what day of the week and which shows are most effective for airline ads.

If it can do all that without seeming creepy or overly intrusive, that’s a huge win for the industry, for marketers and for consumers, who will no longer be bombarded with repetitive or irrelevant ads.

Tags: , , , ,