AI isn’t just a future dream for local news organizations. Its early applications have already arrived, and wider adoption and lower costs will mean an increased presence in a range of roles, experts say.
LAS VEGAS — The early waves of artificial intelligence are already lapping at the shores of local TV news.
For now, AI is largely confined to the realm of metadata and closed captioning among larger, international media companies, according to Ethan Dreilinger, solutions engineer for Watson Media and IBM Cloud Video. But the next wave could see more AI functionality — and affordability — soon enough.
Dreilinger told NAB Show attendees here on Sunday that Watson’s primary role at present is retrieving metadata within a media organization’s video — essentially a search and discovery operation.
Say, for instance, tags have been misspelled in videos long languishing in a media company’s asset management system. AI can root in, correct the misspellings and make the video more retrievable for wider future use.
“It’s really unlocking the metadata to make that video actionable to you,” Dreilinger said.
AI can also make closed captioning an automated part of the news workflow so it no longer needs to be sent out.
But there are other manifestations of AI and machine learning that are starting to appear: conversational interfaces in the form of chatbots at places like The Guardian and Quartz, for instance, and real-time bidding optimization, Omar Karim, head of engineering for Frankly, pointed out.
And Watson is also currently using AI for real-time highlight clipping. At the recent Masters tournament, for instance, it could isolate clips based on crowd noise and the golfers’ own reactions to their shots, locating clips based on the strength of the reactions they generate.
Looking ahead a bit further, Randa Minkarah, COO of Transform, sees AI as helping to create a more immersive experience for a broadcaster’s consumers.
“You stop repurposing one story on television to video on your website and you start to really craft it,” she said.
That means that AI could recognize ingested video content and essentially offer up an “optimized” story that a reporter could then decide whether or not to follow.
Such a machine intervention wouldn’t be a substitute to good journalism, “but we can enhance that and get there faster with more relevant content to all of our viewers,” Minkarah said.
The chief impediment to wider AI adoption right now remains its cost, though Dreilinger said that as tools are perfected, those costs increasingly come down.
And once those costs are lower, he says media will begin to adopt AI by seeing the value of reusing content that they already own that can be reused.
So how long will that be? According to Minkarah, the more progressive station groups will get on board in the next couple of years.
“And it will really take off in the next five,” she said.
Read all of TVNewsCheck‘s NAB 2018 news here.