Videoage International April 2019

24 April 2019 V I D E O A G E AI in the TV Industry (Continued from Cover) (Continued on Page 26) 30 percent of advertisers and publishers are not clear about how artificial intelligence can be used in their work. That’s no surprise given how nascent the tech- nology is in its development in traditional me- dia and broadcasting circles. With that in mind, here’s a look at how AI can transform not just the production of media content, but its distribution and marketing, as well. Famed television producer Mark Burnett once remarked that for every televised hour of the hit reality drama The Apprentice , as many as 300 hours of footage was filmed. Needless to say, that’s a tremendous amount of content to whittle down in the editing room! Post-production staff everywhere can relate as they typically spend hour upon hour cutting, tag- ging, and archiving clips for later use. It’s precisely these sorts of mundane and te- dious tasks that are a perfect fit for AI: they’re highly repetitive, formulaic, and require almost no critical thinking. A recent study by Accenture predicts that the application of AI technologies to these business tasks will increase labor produc- tivity by as much as 40 percent! Most importantly, it will free up human staff to focus on the creative work that drives value for television productions, work that only they can do. A start-up called Clarifai is already using artifi- cial intelligence to drastically cut down the bur- den on video editors. The company and its partner, Vintage Cloud, have launched a platform that uses algorithms and computer vision technology to recognize objects that appear in film and catego- rize them accordingly. What would take humans dozens of hours to sort through takes the machine a fraction of the time, with the added benefit of being more consistent in its classification than a team of humans might be. Taken a step further, artificial intelligence al- gorithms could be used to suggest footage fitting a certain description so editors don’t have to go hunting around for the appropriate content. For example, the producer of a nightly news program could simply instruct the algorithm to find the best footage of a busy shopping mall in Jakarta in the afternoon during the holidays in which there is a major department store in the background that was filmed in the past six months. In sec- onds, the AI assistant would queue up everything matching that description, saving the producer valuable time on deadline that can be used to both improve the quality of the finished piece and free up time for him to focus on other segments. Artificial intelligence can also be helpful in digging up copyright-free images and footage so that producers don’t accidentally run afoul of copyright law while quickly packaging content together in order to meet tight deadlines. Speech recognition and translation is another area in which machine learning algorithms can be a valuable tool in film and television production. The accuracy and speed of speech-to-text tran- scription algorithms has vastly improved to the point where it can reliably handle one of the most time-consuming tasks in content production. A start-up called Fano Labs is training machine learning software to recognize even complicated accents and dialects with remarkable accuracy using natural language processing algorithms and neural networks. traction can help a television program slide into a more favorable time slot where its audience will grow even further. Audience engagement is one thing television net- work executives crave. And when it comes to figur- ing out what makes viewers tick, Netflix is the gold standard. The video streaming titan’s secret sauce is its content recommendation engine, a proprietary mix of machine learning algorithms that can pre- dict the programs a viewer can’t help but binge on. While Netflix has a commanding lead in the space today, there’s plenty of opportunity for TVnetworks to apply their own artificial intelligence solutions to drive audience engagement. Wattpad is one content platform already using AI to great effect for the printed word. Its “Story DNA” machine learning algorithm sorts through the works of more than 70million users worldwide writing in more than 50 different languages to recommend content to readers. The best part about this algorithm-driven solution is that it continues to improve withmore data at its disposal and the right guiding inputs, making it a long-term investment that pays bigger dividends over time. “The machine over time learns from our content repository. But we also use, for example, some classic books,” Wattpad’s chief executive Allen Lau told Beyond Innovation ’s Anthony Lacavera. “We send some amazing classic books to the machine [to instruct it that] ‘Hey, this is amazing, let’s learn from this,’ and the machine over time can pick up the patterns, pick up the writing style, pick up the other signals and find very similar content. That’s how the machine becomes better and better.” Content customization platforms like Wattpad can empower TV networks by delivering an automated way of tailoring programming to individual viewers. This is at a timewhen audiences are demanding personalization and the freedom to watch what they want, when they want. As the AI gathers data on what viewers want to see, network executives can make smarter decisions about which programs to purchase, how much to pay for them, and how much they can chargeadvertisers tobroadcast during commercial breaks. The information gathered is incredibly granular, and as a result, it can be leveraged across multiple mediums and can be used to monetize hyper-specific portions of programming. “It requires a lot of different components,” the company’s chief executive, Miles Wen, told Be- yond Innovation co-host Anthony Lacavera dur- ing a recent interview that aired on Bloomberg Television. “There’s one particular model that’s detecting what dialect you’re using, and there are some other engines that will try to recognize what you’re saying, and there’s another engine behind it that [determines] how to write it out. Machine learning is never a single piece. It’s al- ways a whole bunch of systems pulling together.” The company works with call centers to help operators understand the clients they’re speak- ing with and also licenses its technology to a maker of smart speakers. The technology is com- pletely applicable to the transcription needed in video content production. Artificial intelligence isn’t only for post-produc- tion; it’s also a useful tool for engaging audiences in the social media age. For many television pro- grams, the fan experience that occurs online and through social media channels is nearly as impor- tant as the show content itself! It can even help ex- tend the life of the program on a network, as was the case when fan outrage on social media caused NBC executives to reverse their decision to cancel the drama Timeless . Viewers are more eager than ever to interact with quality content and the peo- ple who create it — AI can help them do so in a way that provides volumes of engaging content for studios and networks. “A huge trend we’re seeing with organizations is that they’re getting a lot braver with choosing creative partners. So rather than just relying on in- fluencers or actors, they’re actually now reaching out and creating with their own customers,” said Emily Forbes, CEO at video production platform Seenit, which helps producers gather high-quali- ty user-generated videos. “We have an algorithm that can push automated feedback such as ‘We love your clip but we can’t exactly hear you’ or ‘Can you not shake the camera so much,’ so the quality of the video we’re getting through our platform is much higher,” she added. Forbes shared on Beyond Innovation that the Seenit platform enables producers to create con- tent on a large scale with production costs that are as much as 80 percent lower than traditional approaches. Yet the ultimate prize for networks is the twofold increase in audience click-through rates on the content that’s created. That added Beyond Innovation co-hosts Anthony Lacavera and Michael Bancroft