Machine learning and more is helping people find TV and film choices suited to their moods
Elly Yates-Roberts |
The days of associating artificial intelligence (AI) with humanoid Asimov-type robots are over; from healthcare to shopping, AI is everywhere.
Conversational AI has quietly changed the game for customer service and what companies can achieve with digital interfaces. Chatbots are ubiquitous – in banking, online shopping and food delivery – but we don’t necessarily register their presence. After all, the idea with these intelligent robotic helpers is that they should make the experience as natural as possible. With advancements in machine learning, natural language processing and advanced text-to-speech, we should be able to achieve conversation so natural that we don’t miss human customer service agents.
It is an exciting time for AI in the media and entertainment space. Virtual reality is pushing the boundaries of world-building and immersive gaming experiences. A few years ago, IBM’s Watson stitched together the movie trailer for the horror movie Morgan, the first to be entirely created by AI. Now, Netflix changes its thumbnails to cater to each viewer, refining what is displayed based on the content being consumed. Like romance? You’ll most likely see couples laughing or looking at each other. Prefer action? You’ll see fighting or a character in action, running or jumping.
But while the algorithms of streaming giants like Netflix and Hulu are advanced and always evolving, they are rather blinkered. In attempting to predict the unpredictable with regards to people’s preferences, they rely overwhelmingly on the history of what people have consumed – on their past watches. Overly preoccupied in this way with analysing past behaviour, they keep returning similar predictions. People, though, have diverse preferences and varying moods. They don’t always want to consume the same kind of content and don’t always like the same genre or the same content creators. They tire of being offered the same fare over and over.
Imagine this: on a quiet Saturday evening, you sit down with your partner to watch a movie or a good TV show. It’s been a hectic week, and you look forward to kicking back with your loved one. But after what feels like hours of back and forth, despite your multiple subscriptions and the veritable cornucopia on offer, you still haven’t been able to pick one that you’re both in the mood to watch. Eventually, you just give up and pick something you’ve already watched a thousand times.
Sounds familiar, doesn’t it?
With unprecedented access to content in this golden age of television and streaming services, ‘what to watch’ has become a common conundrum.
The good news is that with advancements in deep learning and cognitive AI, and rich media chat platforms, perceptive bots capable of picking up context cues and learning from conversation can now help you figure out what to watch. We now have the capability to develop an affable, intelligent bot with the sophistication and adaptability to help you narrow down your broad preferences.
If all you have is a vague preference for something funny, for instance, you can type in ‘funny movie.’ Our helpful bot will return a large variety of movies that can be classified as funny, ranging from mildly humorous to laugh-out-loud hilarious, from slapstick and romantic comedies to subtle black comedy and political satire. You’ll be able to peruse this wide selection and decide what appeals to you in the moment.
Of course, the bot is capable of much more. It holds repositories in multiple languages and can show you a selection from more than one language if you are multi-lingual. It can show you recommendations and suggestions based on past choices, switch languages if you don’t prefer English, ask for clarification if it’s not sure what you’re looking for and learn from experience. It can grow with feedback, picking up vocabulary and understanding a larger number of requests every day, making and storing new connections, and constantly analysing trends and conversation flows.
The bot also has a persona of her own. She can quote dialogues and make suggestions for fan favourites. If you can’t find what you’re looking for, she’ll suggest an alternative. She’ll even ask if you want to check out any of these movies that have been generating buzz, or if you want to give one of these Oscar contenders a shot. She will even try to broaden your horizons by suggesting different genres, edgy and experimental TV, and avant-garde movies. Like any good librarian, she’s not only here to help you find something that you’ll enjoy; she’s here to help you find something you didn’t know you’d enjoy.
Nirvik Mitter is the manager of business development and global marketing at PopcornApps, a Microsoft partner