The current pace of development in AI is remarkable, achieving milestones that were considered many years away
Artificial Intelligence (AI) is suddenly all the rage, with the potential to assist in many areas of the industry from network diagnostics to edit assembly. The BroadcastBridge gets under the skin of AI to ask how useful is machine learning today in media today. Steve Plunkett, CTO, Broadcast and Media Services, Ericsson share his thoughts
How can AI and machine learning be used to help streamline the back-end office or under the hood network diagnostics, crunch big data and / or aid better decision making?
Steve Plunkett (Ericsson): Machine Intelligence provides an important new toolkit to business, with applications ranging from insight (how systems are actually being used) to augmentation (assisting human effort and oversight) and automation (i.e. autonomous system monitoring and intervention). To be useful however, you need to first of all collect lots of data, know how to use it and have a purpose or strategy in place.
Away from the hype, are there real-world examples of AI being incorporated into TV specific platforms and workflows to achieve particular goals (in particular monetization) today?
SP: As an industry we are quite immature in the use of data driven insight and processes but a lot of experimentation now seems to be underway. There are, however, big outliers such as Netflix and Amazon who use data extensively. Examples include rethinking the QC process by analysing errors found in manual and auto QC activity to identifying higher risk material (based on the supplier, technical properties etc) where more time can be spent rather than a uniform approach to all content.
How can AI/ML impact the creative or editorial process of a production today?
SP: Advances in computer vision can allow machine intelligence to automatically log and annotate material, in real-time, which can provide both augmented editorial decisions or automated ones. Object tracking and identification can be used in many situations to save time. The creative process is more art than science so AI/ML should play a supporting role rather than a leading one.
What are the current limits of AI/ML in media?
SP: There are a few fundamental problems today – media applications and infrastructure, in general, don’t emit enough useful data and their control systems and interfaces are designed for people rather than machines. Without enough data, machine learning can’t be effective and unless we can control systems with machine friendly interfaces (such as APIs) then we limit our ability to increase automation and insight.
What should media companies be doing to incorporate AI/ML into their operations?
SP: They should spend time learning about the basics of machine intelligence and how it might help them improve business decisions and performance. They should then experiment to find practical implementations that work for them. This is best achieved with external help and depending upon the results and related strategic decisions, they will probably need to hire in experts in data science, data engineering and so on. One word of caution – embracing data based insights and operations requires a mindset change, ongoing investment and can be met with significant cultural resistance; simply hiring a few propeller heads into a new data science department will not create a data driven business.
What will AI be able to do in the near future that it can’t do now?
SP: The current pace of development in AI is remarkable, achieving milestones that were considered many years away, and there is a particular focus on computer vision, speech recognition and natural language processing. In combination with large scale data collection and analysis we could see AI take on some of the operational roles in media production and distribution that are beyond its capabilities today. We will also see a significant increase in the consumer facing role of AI in content discovery and playback – we see the beginnings of this already with digital agents such as Amazon’s Alexa appearing on TVs and in our lives more generally.
You might also like...
Last year, more than 400 original series hit the small screen. With no signs of slowing; predictions suggest 2017 could beat records, with over 500 programmes launched. It’s now possible for companies to make video easily accessible online and with the widespread a…
Competition in pay TV will drive rapid growth in the video analytics market to reach $3.7 billion by 2022, more than double the $1.8 billion of 2017.
Extreme Networks is the official Wi-Fi and Wi-Fi Analytics provider of Super Bowl LI.
Metadata is well known to hold the keys to good user experiences by making content readily searchable and enabling more compelling or relevant recommendations, but has been held back by limited depth and need for laborious manual production.
Attendees to Society of Motion Picture and Television Engineers (SMPTE) conference might conclude that Machine Learning (ML) and Artificial Intelligence (AI) have progressed way beyond hype as they start to enter just about every aspect of video production and distribution.