When graduates across multiple universities recently started booing their commencement speakers for praising AI, Brian Morrissey saw more than a group of Gen Zs getting angry. He saw the relationship between publishers and their future audiences at risk.

Speaking at WordPress VIP’s recent Media Product Forum in New York, Brian, the host of The Rebooting Show, said publishers are at an unusual crossroads. They need to use AI strategically to improve operations and attract readers, while avoiding alienating employees and damaging their reputations as standard-bearers of objective journalism. 

It’s the difference between inherited trust and earned trust.

“A lot of publishers are operating on institutional credibility…I think everyone out there, whether you’re an institutional publisher or an individual, has to earn that trust. And I think when creating these products, the important thing is not just to keep the human in the loop, but for the human to be central.”

— Brian Morrissey, Founder, The Rebooting

Protecting editorial trust in an AI-driven world

Morrissey cited an NBC News poll that showed only 26% of Americans have positive feelings about AI, compared with 46% who hold negative views. That means they may already be approaching publishers more skeptically as AI becomes part of how journalism is produced and distributed. 

Among employees, meanwhile, he harkened back to “Forditis,” when automotive employees at companies like Ford Motor Co. felt a loss of agency as they moved from directly making parts to simply assembling them on a line.

“Humans ultimately gravitate to other humans, so I think when building (media) products, it’s going to be critically important to have human empathy embedded everywhere.”

— Brian Morrissey, Founder, The Rebooting

Optimizing editorial workflow: doing more without breaking what works

Publishers like Breaking Media, which operates a stable of B2B titles in verticals such as legal, finance, and health, are showing how to strike the right balance between pursuing innovation and preserving trust.

Breaking Media primarily uses AI to improve internal development efficiency, according to their CEO, John Lerner, including front-end design and back-end coding. This has reduced the time required to spin up microsites from one week to a few hours, he said.

AI isn’t being used to write articles, though he noted that publishers need to take into account the “agent experience” (AX), or how well large language models (LLMs) and AI search tools can understand and cite content.

Lerner said Breaking News wants to stay on top of the opportunities AI offers, but in a way that leverages cross-functional expertise and delivers real value.

“We’ve created a couple of skunk works groups in the company, usually with one business person and one editor, and trying to solve a single problem,” he said.

Jess Awtry, Vice President of Digital Strategy and Communications at Pew Research, suggested that Breaking Media is wise to tread carefully in using AI. Pew data shows four-in-ten (41%) U.S. adults believe AI would do a worse job writing a news story than a human journalist, with 66% highly concerned about AI spreading inaccurate information.

“If there is one good thing in all this AI news, it is that human expertise is maybe our last moat…(Audiences) don’t care if you’re charismatic. They don’t care if you’re popular. They don’t care about your follower accounts. They want you to be honest, and they want you to be authentic, and they want you to be smart about the subjects you cover.”

— Jess Awtry, Vice President of Digital Strategy and Communications, Pew Research

They also want publishers to help them connect quickly to accurate, trustworthy information. Yahoo Senior Vice President and General Manager, Kat Downs Mulder, said her company was tackling that mandate with the launch of Yahoo Scout, a search tool designed to minimize the gap between the convenience of an automated answer and the accuracy of verified content.

“Our focus is really on publisher friendliness,” she said. “So lots of links in the answers, transparency in the answers, and really citing sources in the answers, which I think is a differentiator against a lot of the other chat-based answer products out there.”

Crain is doing something similar with its “Crain Answers” product, said CMO Veebha Mehta, providing summaries to search queries while leveraging the brand equity of the company’s well-established journals. “This is as much about learning as it is serving audiences,” she added.

“We’re also getting intelligence from it as to what people are looking for,” she said. “That helps shape our agenda as a publisher.”

Acquisition beyond search: building sustainable growth

One of the surest signs publishers have succeeded in earning audiences’ trust is when they convert from casual readers into paying subscribers. Despite its reputation for offering a humorous take on the news, this is no joke at The Onion, which launched a membership program two years ago and has even gone back into print.

According to the brand’s Chief Product Officer Danielle Strle, charging $99 for a membership allows the company to maintain margins for discounts and incentives like tote bags. More importantly, however, this model fosters a spirit of genuine community rather than a one-way conversation from a publisher.

“The goal is always to make our subscriber experience and our member experience such that they have the tools to become our greatest advocates, and to share the joy of being a member of our secret society.”

— Danielle Strle, Chief Product Officer, The Onion

People Inc. is taking a similar tack in taking its brands into new territories, whether that’s launching People Magazine’s mobile app last year or establishing memberships across other brands. Chief Product Officer, Adam McLean, called its approach an “inversion model” that has led to 19 projects currently underway.

“It’s a big shift for the product and engineering org to go from, ‘We’re building the spaceship, and everyone’s on the same spaceship,’ to developing a bunch of small things that are going in a ton of different directions,” he said. “But publishers are brands, and they need to be able to express themselves in different ways.”

The New York Post has been on a similar trajectory with its own app and expansion into events and new geographies with the launch of The California Post. Ariscelle Novicio, the company’s CTO and Senior Vice President of Digital Strategy, said the key has been gathering first-party data to target specific audiences, such as entertainment industry insiders, and to support the “for your consideration” advertising market.

The other best practice is to keep a watchful eye on performance and make adjustments as necessary. Amid the launch of Page Six Hollywood, for instance, Novicio said the team noticed they weren’t seeing the expected volume of signups. They quickly realized that registering was complicating the journey.

“One of the things I’ve learned about our audience is that they’re often reading about themselves,” she said. “If I have some high-level executive wanting to read a story about themselves, they’re wondering why they have to fill out seven fields first.”

Since then, new visitors can simply enter their email to get started, then follow a more thorough onboarding process separate from the reading experience. This has significantly improved conversion rates, she added.

As more publishers ramp up their AI efforts and add new revenue streams, many of the experts gathered at the Media Product Forum agreed they are transitioning their operational focus from tracking pageviews to analyzing metrics such as Average Revenue Per User (ARPU) and Lifetime Value (LTV). This can require additional data, better measurement tools, and a willingness to adapt to new processes.

Morrissey suggested this is par for the course if media organizations not only want to remain trusted institutions but also be profitable.

“Most publishers are operating an old business while they’re building a new business…It’s really difficult to do both.”

— Brian Morrissey, Founder, The Rebooting