The new packages of streaming media made by Disney, Netflix, HBO, Comcast, and Apple will be launched in succession this month. It’s a coincidence that they all use AI technology to assist their strategic victory. Smart contents generated and updated in an intelligent manner will be a competitive edge of streaming media.
Undoubtedly, smart contents generated by customized software based on demands will be the next tide based on the trend of mainstream AI visual technology. The spread of streaming media has accelerated the maturity of AI-based content generation technology, making it an integrated smart content industrial chain.
Smart contents imply that contents-based experiences are subject to users who watch, read or listen to such contents, and the contents vary with users’ habits. AI technology serves as the basis. To be specific, new machine learning technology, user interface and automatic contents generation technology make story-based personalization, algorithm-based news source, and personalized recommendations the foundation of storytellers. As a result, both news articles and real-time videos will be popular programs in streaming media services.
So far, we have seen some pioneers in this field.
Some measures have been available for the domain of AI autonomous reproduction. This week, Flatiron Institute announced that its AI neural network can successfully establish a 3D simulation of the universe within milliseconds, while it will take conventional supercomputing methods days or weeks to do so. Shirley Ho, leader of the project, said that: “The model can deduce the entire life of an elephant besides recognizing cats and dogs.” In brief, AI neural network can be used for reappearing a phenomenon or a life cycle for things ranging from small incidents (e.g. molecular clouds collapse and become fixed stars) to giant systems (e.g. the whole universe), displaying complete accuracy and resolution.
A science fiction movie based on AI scenarios will be released. According to the Entertainment Weekly, the “AI training sets” and “data inference” provided by Incite, a high-tech company, will be used for the Season 3 of Westworld to design roles, plots, and provide suggestions for its shooting as well as solutions including climate change and story lines of roles’ career growth. According to the producer, it will be a quiz game and a science fiction movie dominated by intelligent computing.
Influenced by AI technology, visual understanding and inference, visual contents generation and 3D reconstruction have displayed amazing effects in the upstream industry of content creativity. The AI content generation technology has been promoting worldwide applications in the midstream industry of production and broadcasting.
Moviebook, a China-based AI company, introduced a smart visual technology system in detail. By using the system, an AI neural network without bias can extract and view one or two static 2D images of scenes by learning machine videos and reestablishing accurate 3D images of such scenes. It should be noted that this network doesn’t make snapshots or complete 3D images but reestablish the visual space viewed and analyzed by all and then produce brand new smart contents that can be used on smart short video content platforms. In addition, this network also allows the system to reestablish 3D objects accurately from a single image, and such objects are applicable to interactive 3D objects such as virtual roles or presenters. According to a technical leader of the company, they are not sure whether the neural network can create visual images in ways specified by content providers so far. However, AGC, the smart image production engine, can solve perspective, occlusion and lighting problems without using man-made engineering. These problems are the pain points of large-scale industrial application of smart contents.
Netflix started using AI to produce automated trailers and titles. It also began utilizing several machine learning algorithms to adjust and improve the presentation of contents in 2017. In that year, the application was launched in more than 100 countries/regions and has over 100 million users. In March 2019, Netflix expanded the method to AI interaction. Content experiences including dialogues and plots can be directly influenced by users’ choices. For example, each episode of its new animation Love, Death & Robots is independent. The plots present four different sequences, and users may make plots, roles and endings in a custom-made manner. In short, plot changes, dialogues and even background music are influenced by such interactive contents choices that are made by intelligent algorithms.
Audioburst, an Israel-based start-up, uses AI technology to extract optimal contents from podcasts and the radio to create new content experiences. Each day, it extracts and indexes millions of audio clips and streaming media clips and uses AI technology (including automatic speech recognition and natural language understanding) to create products such as searchable AV libraries, personalized audio stories and news abstracts, and instant messages. In this way, users can listen to any podcast that they are interested in or any part of other AV programs instead of spending 30 minutes or more listening to a whole lengthy program. So far, the company has raised USD 25 million from Dentsu, Hyundai Motor, and Samsung as its strategic fund.
More importantly, under legal and moral constrains, people-oriented approaches and sustainable values are at the core of smart contents creation. Tik Tok, YouTube, Facebook, Netflix and Spotify’s users are increasing rapidly, and algorithms will recombine relevant contents based on users’ interests. However, you may not know that such contents are made by machines or from smart contents. That has happened, and we will see the new forms of smart contents. User interface, machine learning technology and contents will be combined in a seamless manner to create personalized content experiences.
Smart contents are the ultimate combination of user experience, AI technology and storytelling. Those pioneers grasping such contents will be the rulers in the future, and the era of smart contents is just around the corner.