Will AI Dominate in Broadcasting in 2019?

Will AI Dominate in Broadcasting in 2019?

Artificial intelligence in Broadcast

In almost every area of modern technology, we’re beginning to see the implementation of AI/ML (artificial intelligence and machine learning). AI/ML made inroads in broadcasting and video production in 2018, but these industries are still a ways out from utilizing this technology to its fullest potential. Whether or not it fully takes off this year, or even the next, is a matter of debate in industry trend reports.

In the last year, the adoption of AI/ML in media and broadcasting jumped from 2% to 13% just between April and September. At the 2018 Masters Tournament, IBM’s AI tech was used to identify and compile highlights based on expressions, gestures, commentary, and on-screen graphics. Video producers could then compile highlight reels and broadcast them in near-real-time to viewers. In this and many other scenarios, AI/ML has the potential to significantly speed up production and workflow. 

There are, however, several reasons that broadcasting may lag behind other industries in adopting and deploying AI/ML. The most immediate concern for many broadcasters is that they can’t effectively monetise it. The cost of implementing it may outweigh its potential advantages. 

AI/ML is also best suited to sifting through large amounts of data. To large broadcasters who have plenty of viewer data to work with, this is AI/ML’s biggest benefit. Contrastly, FTA (free-to-air) broadcasters who are moving into OTT often don’t have enough viewer data for AI/ML to be accurate or useful. These companies need to build out their databases beforehand, which requires a data management team, and will inevitably slow their adoption of new technology. 

Finally, the broadcast industry tends to be cautious. AI/ML will undoubtedly see more interest and adoption this year, but broadcasters will need to see significant ROI before an industry-defining change is made. More likely, AI/ML will see moderate experimentation, while transitioning to IP workflows continues to be the priority in 2019.