How to Use AI in Video Workflows
The key to understanding how to best use artificial intelligence in video workflows is knowing when to implement it for the appropriate applications. Ethan Dreilinger, Client Solutions Engineer, IBM Watson Advertising and The Weather Company, Carlos Hernandez, Chief Revenue Officer, SSIMWAVE, and Gordon Brooks, Executive Chairman and CEO, Zixi, discuss the distinctions between automation and AI, along with how AI can be applied to automation to help maximize and streamline production workflows.
“When I talk to media companies about applying AI, I start at the bottom of a workflow and take those mundane tasks and start to think about how you can express them with AI,” Dreilinger says. He believes that letting machines analyze data is the ideal approach, while people such as Producers should handle creative content and cognitive-oriented tasks. “I think in terms of the bottom rung, really start low and build on top of that and take those low-lying tasks and start to automate them,” he says.
Hernandez asks where Dreilinger draws the distinction between automation and AI. “Not all automation is AI-driven,” he says. “So where would you say it makes sense?”
“AI can feed automation, certainly, but automation just kind of runs on its own,” Dreilinger says. “We have some products in our stack where you can go into an app and the app will start to understand what your preferences are because of AI.” This application of AI can be very useful for end users. “'That should go above this because that person likes pollen count more than precipitation forecast,’” he cites as an example.
Hernandez notes that automation and AI are often confused in the market. “We believe that in some cases, machine learning processes make sense, but in others, knowledge-based [uses] are the most appropriate. The key about artificial intelligence is knowing where it has value and where it doesn't," he says.
“It's a tool, like anything else, right?” Dreilinger says. “When you do a home repair job and you need a Phillips head screwdriver, you don't pull out a flathead screwdriver, you get your Phillips head screwdriver. It's a tool to use and you need to know when to use it.”
Gordon Brooks concurs and sums up these points. “It goes through the concept of advanced analytics. It's just one of the tools in that toolbox of doing those advanced analytics, which then drives automation and other things. So, I agree.”
Learn more about video workflows at Streaming Media West 2022.
Watch full-session videos from Streaming Media East 2022.
Rema Morgan-Aluko talks about the ways her teams at Fandango, Vudu, and Rotten Tomatoes have developed unique methods to consolidate both short-form and long-form workflows across platforms while also optimizing speed, quality, and costs.
Zixi has improved live video workflows through their specialized Software-Defined Video Platform, which uses dynamic machine learning and an automated analytics approach to Root Cause Analysis to assist with faster team problem-solving collaboration
Some streaming pros say cloud production has made 5-10 years of progress in the two years since the pandemic shook live production to its core, but when will cloud offer the sort of no-latency communication, preview, and replay producers expect and rely on in on-prem workflows? And will it require a paradigm shift in the way producers think about hardware purchases and usage? Live X's Corey Behnke, CNN's Ben Ratner, LiveU's Mike Savello, and Signiant's Jon Finegold discuss the challenges of cloud migration and the current state of cloud in this clip from Streaming Media East 2022.
Zixi VP of Business Development Eric Bolten discusses some of the prevailing misconceptions about AI/ML video workflow automation in this clip from Streaming Media Connect 2022.
Companies and Suppliers Mentioned