NAB
Now more than ever, broadcasters are being asked to do more and more with the same or even fewer resources, but artificial intelligence could ride to the rescue. If used correctly — that is to say with caution to ethics and copyright — AI can create a production optimization opportunity for the industry, a panel of executives at NAB Show explained in a session titled “Using AI as a Creative Content Tool.” Watch the full conversation in the video below or read on for the highlights.
article here
Tegna operates 64 television stations in 51 US markets, reaching approximately 39% of all television households nationwide. It invited its journalists to suggest ways in which AI might improve their efficiency. From that the company has begun deploying AI.
Kurt Rao, Tegna SVP and CTO, explained, “Most newsrooms get thousands of emails a day but with limited number of reporters so the question we might ask ourselves is, how do you get signal from noise? So we’re using AI to really distil all of the input coming in.”
Another example is being able to create a recommendation engine that will target different lists of news stories to its audience, built from an AI scouring video and transcripts of its coverage of school board meetings and government meetings.
“The next step, which we’re working on right now, is you can literally highlight text in the transcript and it edits a video for you. These kinds of use cases will really expedite the production process. My hope is if this person is capable of doing two to three stories a day, we might be able to double their output.”
Tegna is wary of letting its staff loose on ChatGPT and similar public AI models, and even more wary of publishing anything on its platform without proper vetting.
“We’re having ethics policy training. We’re very careful about what actually makes it to platform. If they do use those tools there are disclaimers put in place, and we have additional reviews of content that may have come out of a AI like process,” Rao said.
Stringr combines original footage sourcing with video editing and management tools in a cloud-based solution — streamlining remote video news production. It is supported by a network of 120,000+ videographers who are stationed across the United States.
Brian McNeill, founder and CPO of Stringr, said, “If you think a typical news production package might cost you between $300 to $2,000, for your two minutes on air, using Stringr that might get down to $100 or $200 because of our cloud-based infrastructure. Using AI, they can get sub $1. This opens up a whole new world. It makes some of the FAST channels make economic sense that didn’t previously make sense.
And this is really just the tip of the iceberg to a whole new world of content that’s going to be either more targeted or even targeted to an audience of one.”
The Weather Company has a partnership with NVIDIA where it uses NVIDIA chips and data centers to power the weather on NVIDIA’s Earth-2 platform for climate change modelling.
But the company is not using GenAI just yet, “because we don’t trust it yet,” said Joe Fiveash, VP of Enterprise Media.
The Weather Company claims to be the world’s most accurate weather forecaster and won’t introduce anything that might dent that. However, AI could be used soon to tailor its licensed weather data packages to TV stations.
“All of the videos that we create are branded to the station where creating content that’s friendly for sponsors is really important,” Fiveash said. “It’s just hard to do when you have a staff of a couple of meteorologists. But now they can do the golf forecast and the regular forecast and the boating forecast in their own station voice [while saving resources on time and cost].”
One of the biggest challenges for news organizations in the age of AI is maintaining their brand integrity. They can’t just throw a story into ChatGPT and publish.
FOX is combatting this with Verify, a tool applied to all FOX content which declares both the source of information and any compensation for use of data.
Melody Hildebrandt, CTO at FOX, said, “We think in the future there’s going to be deliberate misinformation and taking information out of context that flood the zone.
“Where you just have so much to sort through that you really need to have the ability for a consumer to essentially rely on the broadcasters or publishers that they trust to help them navigate that information space.
“One of our core hypotheses is that broadcasters like us are going to be essential. Brands are going to be more important to help consumers navigate that information space. But we need to be able to bind digital content to that real world publisher.”
Hildebrandt added, “I think as a community of publishers, local news, producers, broadcasters should be opinionated about the architecture of how these large language models see our content.”
She described how Verify would work, saying it was essential for consumers to be able to validate that what they see online came from a source that they trust.
“As an industry we should be banding together against Big Tech,” she said. “Rather than have Big Tech impose on to us, we can actually all be at the table and say this is how content should be consumed by these models. We can all have more confidence in the use of those models if we have this transparency into how they were trained.”
Mike Palmer, AVP of Advanced Technology and Media Management at Sinclair, explained why the media conglomerate had joined the Coalition for Content Provenance, a similar content verification initiative.
Other members include Adobe, BBC, Google, Intel, Microsoft, Sony, and vendors like Telestream.
“There’s a string of signed certificates going back to the base URL that says this has not been faked,” said Palmer. “This is actually who you think it is. And you can look at the chain all the way up from the GPS coordinates baked in metadata into the camera in the field.”
No comments:
Post a Comment