News
ePublishing’s New Report Explores the State of AI in the Publishing Industry
September 17, 2024
Imagine that your company has a promising and talented new hire. They’re versatile, resourceful and have already racked up some impressive wins. But they’re also a little bit of a maverick. They don’t always follow instructions, they have a tendency to cut corners — and you suspect them of sometimes just making things up when it suits their purpose.
So you find yourself in a place where you have every reason to hope that as time goes by, they’ll become one of your most valuable players — but for the time being you need to keep a close eye on everything they do, assign them a more experienced partner to double-check their work, and keep them on a short leash to limit your liability.
This might sound like the tried-and-true premise for a detective show on TV. But it’s also the situation that many publishers say they find themselves in regarding the promise of Artificial Intelligence (AI).
To get a better sense of how publishers are managing these challenges, ePublishing decided to try the direct approach: We asked them.
We found that they’re aware of both the potential rewards and the very real hazards of working with AI — and they’ve adopted a cautious approach to move forward wisely.
Access now: State of AI in Publishing
Who Did We Talk To, and When?
Our goal in conducting this survey was to gain a deeper understanding of publishers’ views of AI. We wanted to uncover the challenges this technology poses to publishers, as well as the opportunities it provides.
Completed in June of 2024, ePublishing’s new report draws from an ePublishing survey of news, B2B, B2C, data and journal publishers and includes insights based on interviews done with publishers across segments. Thank you to everyone who participated.
Here are our key takeaways:
How Are Publishers Using AI?
It’s important to underscore that we didn’t find a uniform approach to AI among publishers: We found different organizations moving at different speeds, and with varying levels of confidence. Some publishers were resistant, but open to functional use cases; others were already heavily leveraging AI to speed up high-volume repetitive processes.
That said, a common theme was that most publishers aren’t ready to hand over the task of writing most content to AI — and certainly not without careful oversight and supervision. Instead, most of those publishers who have begun to implement AI are using it in a supporting role: to assist with tasks like generating headlines (38%) and transcribing interviews (32%), or to support the writing process with research, outlines and topic ideas (36%).
Writing promotional and marketing copy (34%) or social media content (32%) are considered less of a risk than writing articles for publication, which only 21% of publishers are ready to try.
Image generation is another area where publishers are taking a cautious approach, with only 27% currently using AI for this purpose. And the vast majority aren’t ready to embrace AI for video creation, with only 9% taking the plunge so far.
Publishers not only want to keep a close eye on any work generated by AI, they don’t consider it ready to accurately evaluate the work of human writers, either — only 13% of publishers find it useful for partial or full editing.
How Is AI Benefitting Publishers?
One of the most-cited benefits was saving time, with “quicker turnaround” and “acceleration of the content creation process.”
Respondents also appreciated the ability of AI to conquer writer’s block and move the process forward, saying that AI “gets us started,” assists with “idea generation” and “can sometimes help with headlines when we aren’t able to come up with anything.”
Some publishers have also found that AI is adept at tasks like writing SEO-optimized headlines, click-worthy email subject lines and short summaries for longer print stories.
Boosts in efficiency, productivity and accuracy were mentioned by multiple respondents. One publisher noted the rewards in terms of removing friction from the publishing process: “It has helped in streamlining certain aspects of our workflow that were tedious, and has given us time back for more important tasks.”
Another publisher told us that the ability to automate low-level production tasks had already paid off in the ability to devote more resources to the process of creating quality content.
Key Concerns: Trust and Authenticity
The consideration that publishers mentioned most boils down to one key word: trust. Work that’s generated by AI needs to be closely vetted for accuracy and relevance. Sourcing of information is a key concern, as well.
Seventy percent of publishers surveyed named editorial integrity as critical. That goes hand in hand with the need to be transparent with readers. An awareness of readers’ mistrust of AI and ongoing questions about the ethics of using AI in journalism were top of mind, as well.
Some publishers are also concerned about the investment needed to get up and running with AI — in terms of the necessary training and skills, the effort of finding the right technology and the costs of implementation.
The cure for this anxiety is a focus on keeping it real: making sure to preserve authenticity in a way that maintains readers’ trust. One respondent wrote of the need to make sure that “our readers know that we are human beings making connections to our followers and subscribers — rather than just copying and pasting what AI provides.”
On a creative level, voice matters. AI-generated writing can have a tendency to feel generic, unnatural, lifeless or lacking in the human touch — so it may need additional work from human beings to fine-tune it, add warmth or personality, and make sure it rises to the necessary standards.
Related Challenges: Fact-Checking, Data Protection, and IP
One news publisher wrote that they weren’t willing to embrace AI for content creation. “AI is simply an algorithm at this point, and cannot replace human minds for unbiased, local news reporting.” They observed that their staff had “impeccable integrity” when it came to sticking with proven facts. For that reason, they felt that editing and proofing AI content for accuracy would be more time-consuming than writing articles from scratch.
Some publishers say they aren’t comfortable using AI for image, video or content generation because “the copyright issues are not yet clear” and the work of AI is “much harder for us to fact check.” One publisher worries that AI could result in publishing “inaccurate information that could have big legal repercussions.”
Most respondents named the risk of publishing inaccurate or misleading content, as well as the danger of intellectual property rights infringement, as top concerns.
One publisher wrote bluntly: “I don’t want to feed it our information.” Publishers naturally worry about the extent to which AI may be “ripping off” their own IP. Also fueling distrust: Some publishers worry that the use of AI in search engine results could adversely hurt their traffic.
For these reasons, some publishers feel that AI should be seen as “not a content generator, just a content optimizer.”
Looking Ahead: Publishers’ Plans for AI
Despite their concerns, most respondents feel that AI is here to stay and will play a growing role in their futures. Only 21% said they didn’t have any plans to expand the use of AI in their organization in the next 12 months.
When asked where they saw the greatest potential for AI use within their organization, publishers focused on editorial workflow applications, with marketing functions running a strong second. Back-end operations like finances were on the radar for some publishers, as well.
One of the greatest areas of expected growth is in the use of AI for recommendations and targeted content on their websites. More than a quarter of respondents intend to implement this in the coming year, as opposed to the 6% who have already done so.
Where Do You See Yourself in Three Years?
When we asked publishers to extend their view three years into the future, there were pessimistic and optimistic assessments of the road ahead.
Some were worried that AI would exacerbate the “dumbing down” of the industry, and wind up causing misinformation to circulate among readers who don’t take the time to vet content for accuracy — resulting in a “tsunami of untrustworthy content” that’s been recycled from poor sources.
On the optimistic side, publishers are hopeful that it will allow reporters and other content creators to put more of their focus on writing, as AI relieves them of repetitive tasks and assists them with research.
Publishers also hope to benefit from AI’s ability to help them with marketing and social media, as well as analyzing reader habits and content consumption to provide useful insights and add value to their products.
One respondent summed things up by saying that the winners of the AI revolution would be those publishers “who can leverage AI to increase productivity, market share and margins.” But they took a realistic view of the challenges involved, adding: “It will be very tough.”
We’ve just scratched the surface here — publishers had lots more to tell us. Read what they had to say, and look at a breakdown of the survey data in the full report, “The State of AI in Publishing,” which is available free for download.
Read CTO Trey Connell’s take on AI in publishing: Artificial Intelligence for Publishers - Friend or Foe?