• AI at the service of the media industry
  • How the media can benefit from AI
  • AI: the throbbing heart of the media industry?
  • How AI transforms the media industry
  • AI is the driving force of modern media, and here’s why

Slowly but steadily, AI is making its way to new industries, transforming them from the inside – adding value, substituting people workers in some areas, and generating new jobs in others. The media industry is no different. Artificial intelligence impacts all the parts of the media delivery chain, from content creation to consumer experience. It is an important tool that allows publishers’ to stand out in a crowded marketplace by an increasingly important feat: delivering the right message to the right users at the right time.

AI is not a completely new concept. Still, it’s been gaining more traction recently due to an exponential increase in computer processing power and storage capabilities, as well as the rate of information exchange on the internet. In this article, we map the possible areas of disruption and provide examples of the impact of AI on the media.

AI for automation

The main promise of AI at the service of the media industry is the automation of some of the routine workflows. Netflix has some lofty claims about it, saying AI allows to save a US$1bn worth of human labor – making its recommendations system better. The overarching goal is to make subscribers happy with the content they’re finding, and reducing customer churn in the process.

On the more technical side, AI can automate some mundane tasks associated with the conversion of data formats, transcoding, extraction of audio and subtitles – to accommodate the content to various devices.

But AI working at the services of the media is not just limited to video. Layoffs among reporters and editors have taken place at Reuters and Bloomberg News, where Cyborg, a piece of AI-supported software, makes sense of financial reports almost in live time. It comes up with news stories, including key information and breaking news.

There are more examples like this. Freshr, a chat bot, prepares a summary of the biggest news of the moment based on the user’s tastes, every morning, in just 5 minutes. It is aimed at 20- to 35-year-olds. It is estimated that automated AI technology accounts for one-third of the content published by Bloomberg News today.

Your Weekly Edition is a personalized newsletter sent out by The New York Times. To deliver a curated, personalized selection of content to NYT readers, the newsletter engine uses both algorithms and human labor. The goal is simple: show users more content they haven’t seen yet.

Better video and audio content discoverability

With the amount of video and audio content available at people’s fingertips, it becomes a real challenge to give people a tool allowing them to search and find the piece of content that’s really relevant – at the very moment.

In the early days of the internet, search engines only had to deal with text. It was relatively easy by today’s standards – typing a few search terms allowed to reduce the number of results.

Today, to allow people better search within video and audio content, AI first has to make sense of the content, and be able to “see” it in the way people do – faces, words, objects. Powered by AI, searches can now be done on images, videos and voice recordings by a combination of a few technologies:

  • Image recognition technologies
  • Machine learning
  • Voice transcription techniques
  • NLP
  • Face, object and place recognition.

Google has been working hard on voice transcription and object recognition technologies for years – e.g., videos hosted by YouTube can take benefit of automatic captioning, which works better by the minute. Google Lens does a great job recognizing everyday objects with its image recognition technology. During the 2019 Made by Google and Pixel 4 event, Google demonstrated a new voice recorder for Pixel 4, which automatically generates voice transcriptions in real time of any recording. Importantly, the chip responsible for the process sits right in the device itself, and no internet connection is required.

Better archiving and metadata creation

Without appropriate metadata, it would be impossible to find a specific object in content that has been produced. This is why it is essential to harness the power of AI to automate metadata creation – when fed with sufficient amounts of data, the process is faster, less expensive, and more precise.

AI helps to automatically generate metadata for video content. It enhances archiving and, in particular, makes the content more discoverable in future searches. Azure Video Indexer, for example, leverages media AI technologies. It recognizes objects, people and voices and extracts the metadata from the video.

The possibilities here are plenty. AI can generate time-coded transcripts, recognize faces, speakers, objects, actions, brands, keywords and even sentiments – for content moderation or censorship purposes.

In addition, thanks to machine learning processes, it is also possible to train the system in order to adapt it to the specific needs of its users. An example would be identifying actors and voices on the fly in a way that allows.

Better content engagement

The extracted metadata can then be used for improving user experience. Accurate speech transcription is the foundation of translation in multiple languages. It can further refine the recommendation algorithms by the objects and people that appear in a video, and automatically create clips with the sections featuring a particular person.

Media companies could originally forge their competitive advantage through properly utilizing two concepts: content and distribution. So now we look at a new competitive advantage: relevancy. AI provides contextual relevance for any individual to connect people to relevant content at the right time and in the right context, it’s all about the experience you provide to consumers.

Better targeting and personalization

AI lends extremely well to applications which require crunching significant amounts of data: analyzing user trends to target the best time to broadcast specific content. It can analyze audiences, automatically generating titles/summaries/illustrations with keywords and hashtags that guarantee greater content visibility, personalized newsletters, customized playlists.

Advances in AI could help sweet spot between customization – i.e., giving the users more of what they like and smart content promotion – suggesting completely new titles that the viewer might enjoy.

The operation of platforms like Amazon and Facebook has, for years, been contingent on good targeting and personalization. Netflix is no different – it tailors its entire home page, promoting specific content to its individual users. Its Meson workflow automation system works based on machine learning (collecting data to evolve constantly) and offers a customized visual (9 versions) on which the user is most likely to click, depending on their usage journey and context. The goal is simple: to find, at the smallest use of resources, the greatest combination of series that could satisfy individual, rather than all users. Algorithms are thus underlying creativity and diversity, rather than standardization.

Better monetization through advertising

Thanks to the integration of AI tools and programmatic advertising technology within media management systems, advertisers are given greater control of all the available media, and deliver ads more precisely to specific user groups.

AI can also help fight ad fraud, i.e., fraudulently representing online impressions, clicks, conversions or data events for financial gain. This means more effective spend and better targeting of all advertising.

Audience analytics

Because AI also guarantees increasing insights into audiences, it can be used for content monetisation – e.g., in advertising and content licensing – and customer retention (viz Netflix).

With more detailed audience data, publishers can run effective customer retention campaigns or feed personalisation algorithms to foster customer loyalty and better match the needs of the viewers – something key to services like Netflix of HBO Go.

Microsoft Azure Data Platform is used to capture data about user interactions with online media, building user profiles (also of anonymous users) that power recommendation engines, personalisation, ad targeting and inform content investments. Netflix and other streaming services use data to make recommendations and develop original content that keeps subscribers streaming.

It’s an opportunity that has never existed before in the history of publishing and media; by analyzing online behaviors and other data, brands can accurately recommend products or links, serve up ads in extremely timely moments, and tighten marketing budgets while increasing conversions.

It’s something that’s already happening on both large and small scales, with businesses of all types and sizes leveraging data to disrupt.

Summary

As evidenced in the examples above, AI can take a bird-eye view on the data and see usage patterns and tendencies. The technology is being utilized by major news sites or platforms like Spotify for regular recommendations (depending on the time of day, weather, etc.) and accurate content curation to appeal to individual palates. But it’s not limited to video. According to Reuters, 59% of media use artificial intelligence to recommend its users other articles.

Also, the automation that AI enables isn’t necessarily about making people redundant. The people-or-machines approach is not exactly the way AI disrupts industries. If anything, one complements the other. AI allows people to do much more by relieving them of the mundane workflows, and allowing them to focus on the areas which benefit more from the human touch. And while some jobs will be obsolete as a result of the AI revolution, there will be many more jobs in many other AI-related areas in the future.

Proper leveraging AI involves seamless integration of people and automated components to do more things, faster, better, and with greater levels of personalization for the consumers.

If you are interested in Artificial Intelligence in video, read one of our previous blog post about Automatic Content Enrichment – our proprietary innovative system enabling real-time description and categorization of video content.