It did not take too long for it to appear in the trade journals - "The Future of Artificial Intelligence in Broadcasting" click here is being approached by TV Technology in an upcoming webinar. What will they discuss? I think there are almost too many obvious points to consider. For today let us just speak on live production.
The automation of non-scripted production would be one of my favorite topics. Today to outfit a control room for a corporate training center you might spend as little as <$50k dollars. Go back to the late 90's when i first started working in production and you were absolutely looking at a minimum of six figures to accomplish the same level of production. Naturally all the advances in FPGA chipsets and networking are to thank/blaim. The number of solutions available today to pull off a watchable live event are vast and inexpensive. So where can progress take us?
The "automation" of today whether it be for production or playout generally depends upon scripts or repetitive tasks. Some technology can be programmed to perform action according to if/then logic scenarios, but this is really not artificial intelligence in my way of thinking.
One of the more recent examples of automating live unscripted production was demonstrated at NAB 2017 by Pixellot. Now this, is more like AI. Pixellot is an Israeli company focused on automated sports production. They already have a footprint in the U.S. in the high-school sports space with folks like PlayOn! sports. Basically imagine a locked camera giving you access to any possible angle and then with algorithms following the action like a camera operator might. Sony demonstrated something similar with 4K technology several years ago.
It is just a matter of time that scene recognition, timing, and framing algorithms will replace producers and directors for many of the lower end productions, and it will all ship in a single box sold by Amazon.
....So let me add a footnote to this blog post. Here is another link about AI in the commercial integration space. -