This article was originally published in the Summer 2018 issue of The Record.
Artificial intelligence (AI) promises to transform the media and entertainment business – impacting everything from content creation to the consumer experience.
“AI will influence all parts of the media value chain, helping content creators to be more creative, helping content editors to be more productive, and helping content consumers to find the content that matches their interests and current situation,” explains Rainer Kellerhals, Microsoft’s Media and Entertainment industry lead for the EMEA region. “It will assist human creativity and human curiosity by taking a lot of the leg work out of finding relevant content, navigating large amounts of content, and re-formatting and re-purposing content.”
Lorenzo Zanni, IABM’s lead research analyst, agrees. “Media companies can leverage AI throughout their content supply chains to automate operations, drive decision-making and personalise the consumer experience,” he says, pointing towards automatic metadata tagging and extraction as an excellent example of one of the most effective use cases of AI today. “Through techniques such as image recognition and speech-to-text transcription, metadata tagging is the most widespread application of AI so far. The metadata automatically created by the AI algorithms can then be used to drive content monetisation strategies.”
And this is just the start. Zanni says that media companies can also use AI to strengthen their predictive capabilities. “For example, AI tools can be used to predict demand to adjust resources (in on-demand cloud models) or to predict possible disruptions in the content supply chain (such as a content supplier failing to meet a deadline). These use cases could bring sizable savings to media companies.”
When it comes to distribution, AI can personalise the consumer experience, driving title recommendations and curating content based on consumer preferences. “This is consistent with the transition from a ‘one-to-many’ to a ‘one-to-one’ model,” Zanni notes.
Despite this potential, Zanni believes that true success to date has only been achieved by the pioneering few. “Although media companies are already using analytics tools to analyse operations and audiences, they are just starting to harness the power of more sophisticated tools such as deep learning algorithms,” he says.
There are a number of reasons for this, most of which revolve around data. In supervised learning algorithms, datasets need to be labelled by humans to train the model, making the process expensive and cumbersome for large datasets.
“The availability of training data is a particular challenge,” explains Kellerhals. “Many AI methods use some sort of machine learning, and in most cases, AI can only be as good as the data which is used to train it.”
“Deep learning algorithms produce the most accurate results only when they are fed with millions of observations,” Zanni adds. “Therefore, media companies need to manage different types of data in a unified manner to power effective AI-driven decision-making. This data includes audience data, operational data and content data (metadata).”
To succeed, media companies need to deploy technologies and implement strategies to gather data at scale. Zanni says that a “data-first” approach is necessary – something that heavy users such as Netflix and some of the niche over-the-top players are adopting. “Most of these companies have moved data processing workflows to the cloud, as this allowed them to scale up resources if needed by the size of the information analysed,” he explains. “This leads not only to what all media companies are looking for at the moment – a better return on investment – but also to a greater responsiveness to market changes. What’s more, as they move to the cloud and establish direct to consumer connections, they’ll be able to gather more data on operations and audiences.”
This is where the Microsoft Azure cloud comes into its own. It offers powerful machine learning, real-time analytics, cognitive and bot capabilities through open APIs, which enables companies operating across the media and entertainment industry to transform their content and audience data into a competitive advantage.
“Azure Video Indexer, for example, builds upon media AI technologies to make it easier to extract metadata from video, including timecoded transcripts, faces, speakers, objects, actions, brands, keywords and sentiments,” explains Kellerhals. “Our audience insights function, meanwhile, uses the Microsoft Azure Data Platform to capture data about user interactions with online media, building user profiles (also of anonymous users) that in turn are used to power recommendation engines, personalisation, ad targeting and inform content investments.”
With these types of solutions, the potential for media and entertainment companies is huge. And, according to Guy Finley, executive director at the Media and Entertainment Services Alliance (MESA), it’s an opportune time to make use of this potential. “As an industry we are beginning to build a direct relationship with the consumer for the first time and we’re already working to automate and digitise existing workflows and processes,” he explains. “For example, when contracts become smart, and we migrate from what was once called a ‘one-button transcode’ to an AI-enabled ‘no-button’ transcode, we will begin to see how fluid this entire process can become. Redundancy will be further reduced in our supply chains by smarter, data-driven systems and AI will ultimately drive the globalisation of our enterprises, enabling production and distribution to meet a growing, but still localised, demand for our content.”
Possibly most significantly, AI will be at the forefront of creativity – the force that ultimately drives the media business. “Artists equipped with an AI-enabled feedback loop based on real-time, consumption metrics will up their creative batting average, which will thus increase production and commercial ROI,” Finley concludes.