Station Groups Ready Up For Gen AI’s Industry Transformation

Sometime ago, the leadership at Graham Media Group sought to determine how much the company’s employees were using generative AI services like ChatGPT to do their work. So they sent out a survey to the entire group. “There was a spectacular response. Massive results,” said Michael Newman, director of transformation, Graham Media Group. “The number of people using AI, even for day-to-day headline writing, initiatives, story drafts, was really surprising.”

The task force Graham put together to tackle issues related to AI realized “how critical it was for us to make sure we had clear guidelines,” Newman said. But at the same time, “we wanted to bottle up that excitement into something that we could share and optimize and use to serve our communities better.”

Newman made his remarks during a TV2025 panel session Wednesday that focused on how station groups are finding ways to cut costs and potentially generate more revenue through the use of generative AI technology. And the panelists also discussed what they’re doing to guard against gen AI usage that could get station groups in big trouble.

Stations have used of older forms of artificial intelligence for years. “We have six local newsrooms, some of them in the smallest markets in the country. And we’ve centered AI around making our jobs easier. Using it for social media and understanding what consumers might be interested in have become commonplace. Captioning is another example. Transcriptions have been standard,” said Colin Benedict, VP of news at Morgan Murphy Media. “We look at generative AI through a similar lens: how do we use it to help our journalists do a better job?”

“The cost of ASR [automated speech recognition] is a fraction of what it costs us to do live captioning,” added James Finch, VP of news services at Gray Television. That said, Finch also noted that gen AI technology needs to be properly trained. The company is keenly conscious that garbled messages created through gen AI could result in FCC complaints.

While tasks like transcriptions have been available through AI programs in the past, generative AI brings the advantages up to a whole new level, explained Philippe Petitpont, CEO/co-founder, Newsbridge. “Transformers can train massive amounts of datasets very, very quickly. This is opening up massive opportunities.” By transformer, Petitpont referred to gen AI’s deep learning architecture.

As an example, Petitpont said that creators of a scripted show could conceivably come up with new ideas for storylines based on gen AI’s ability absorb and analyze what’s been discussed on social media. And it could also generate 10 to 15 teasers that would be targeted to different audiences.

Ahron Young, CEO, managing editor of Ticker, said his team also uses ChatGPT to develop questions for presenters and to form 300-word summaries of interviews his service has produced. Ticker also uses gen AI to identify potential native advertising clients on LinkedIn who might be interested in being featured in Ticker interviews. With gen AI, team members can produce customized, six-page PowerPoint presentations that use the potential client’s graphics and color scheme as well as Ticker’s own.

“And it takes five seconds,” Young said. “I’m not going to say that we send out the first version.” But it helps the marketing and sales team members imagine what a polished presentation could look like, he explained. And that speeds up the process.

The session’s moderator, TVNewsCheck Editor Michael Depp, noted that there has been no small amount of worry about the negative implications of gen AI. In response, Graham’s Newman said: “I think a lot of the fears are well-founded.” Among the concerns are news stories entirely created by gen AI that could have been produced by humans in newsrooms. And it’s difficult to sort through stories produced solely with gen AI and what’s been produced with actual journalists.

“Losing the human content is really an important piece. AI does provide really good first drafts of items, but a human needs to be there to tell the critical pieces of the story,” Newman said. “The other thing to keep in mind is a lot of us cut our teeth rewriting stories and turning them into a narrative format. And AI does that really well.”

“Our quest for efficiencies can’t override our service to the public,” added Gray’s Finch. “If we’re not producing a product that the public needs, then what are we doing?”

As with Graham, Gray formed a task force to monitor how the staff is using gen AI, the technology’s developments, as well as ways to communicate with the public about how the company is using it, Finch said. The task force meets quarterly and includes the leaders from critical areas of the company, including news, engineering, sales and marketing. And the company has also placed information for the public on station sites that explains that all the company’s journalism is created by humans, with links to another page that goes further into depth about Gray’s AI policies.

Gray serves 113 markets. Keeping track of what the huge workforce is doing with gen AI makes that task force essential. “The more people you have, the more concern you have that someone is going to do something to violate copyright or breech trust with the audience,” Finch noted.

Read more coverage of TV2025 here.