Digital Innovation Digital Publishing Top Stories
5 mins read

Unlocking the power of AI: 6 lessons from AP for publishers

Next-generation AI chatbots are like London buses. You wait ages for one, and then three come along at once (ChatGPT, Google’s Bard AI and Microsoft’s new-look Bing).

This has several significant implications for publishers, potentially shaking up everything from search to advertising and content generation. Although the initial hype has been met with a reality check, this is still the hottest thing in tech and media right now. Moreover, it’s worth remembering that AI is already integral to the work of many publishers and media companies. In fact, it has been for some time

In 2014, the Associated Press (AP) was one of the first news organizations to adopt AI. Since then, they have continued to be a trailblazer, using this technology for story gathering and content creation, as well as the distribution of material. 

What can other media companies learn from AP’s expertise with this tech? To help unpack this, we spoke to Aimee Rinehart, program manager for The Associated Press’s Local News AI initiative, about some of the lessons their work has yielded. Although Rinehart’s work is focused on AI in local news and journalism, these efforts offer lessons for publishers of all shapes and sizes. 

Here are six things publishers need to know:

1. Companies want AI to remove tedious tasks

As part of their local newsroom project, AP surveyed nearly 200 newsrooms across all 50 U.S. states. This engagement revealed a strong desire to use AI to automate routine and often repetitive (sometimes monotonous) tasks. “They see this [tech] as a potential for help,” Rinehart says.

According to survey respondents, automating transcription topped their automation wish list, followed by social media content creation. Respondents also expressed interest in automating structured data-based stories, such as earnings reports, weather updates and sports game recaps. 

Earnings reports, of course, are an area where AP pioneered use of AI technology. “AP went from doing 300 earnings reports to 3,000 earnings reports [a quarter],” Rinehart says. Alongside this tenfold increase,” some businesses said that they saw an uptick in trading on them, because they were they were now searchable,” she adds.

Image: Results from AP’s artificial intelligence readiness survey in November 2021, via AP

2. Many publishers are wary of adopting AI technology

Although many media companies have, rightly, been accused of suffering from shiny object syndrome, Rinehart paints a slightly different picture. 

When it comes to AI, local newsrooms “are wary of adopting another technology,” she says. The reason for this is that many newsrooms have been burned before. 

As Rinehart explains, “they’ve adopted other technologies that get bought and sold, changed, buried… everybody gets trained on and becomes dependent on them, and then they go away.” Meta’s proposed closure of CrowdTangle and Twitter’s tweaking of their API, are just two recent examples of why these newsroom fears are justified. 

3. AI risks creating a new digital divide

“There is a significant gap between large and small news organizations in terms of how widely AI and automation technologies are used,” revealed AP in a report on their initial work with local newsrooms. 

As Rinehart illuminates, money is often a big factor, especially for smaller outlets. “We had one newsroom we spoke with who said we just dropped our Hootsuite license because it was $100 a month. Another newsroom said we just dropped LexisNexis. And so, newsrooms are really looking at these monthly recurring charges. And really, the $100 threshold is, it’s quite high for many small newsrooms.”

Alongside this, there’s also a concern that the benefits of AI have been “overpromised,” Rinehart notes.

Smaller newsrooms are doing more with less. And although it can potentially benefit them, adding AI into the mix may be just one thing too many. “it’s already kind of overwhelming,” she observes. 

4. Take time to learn about AI and its potential

Toward the end our discussion, I asked how leaders of a two person, or 200-person, newsroom should be responding to AI. 

“First and foremost, to not panic,” Rinehart answered immediately. “Be aware that we are at a [start of] the third digital wave. And it is better to begin to be conversant in this space than to avoid it. Because it’s not going away.” 

“You have to take the time to learn it a little bit,” she adds. “You have to take the time to learn what’s possible from it, and to approach it with a with a curiosity.”

To help industry leaders to do this, AP published a report based on their survey results, and more than two dozen in-depth interviews with local news leaders. Print, radio, television and digital players are all featured, alongside commercial and nonprofit operations.

The report was accompanied by an online training course for leaders, which features insights from leaders at The New York Times, Vox and AP, as well as academics based at Northwestern, NYU and elsewhere. Materials, including tip sheets and videos featuring demos and discussions with industry leaders and practitioners, can be found for free on the AP website

Image: Screenshot of areas covered by AP’s online program on AI

5. Start with an end goal in mind

“This space has really forced a lot of journalists to be technologists, and with really bad results,” she says. “We need help in every direction. Where do we start?” Rinehart asks. 

She recommends that leaders identify annoyances and then look to see how they can potentially use AI – and the myriad of different uses and functions this umbrella term encompasses – to tackle them. 

For a long time, “I wasn’t that impressed with AI,” Rinehart admits, but “with tools like DALL·E (a new AI system that can create realistic images and art from a description in natural language) and GBPT3 [it] feels like magic. There’s a glee to it. There’s a joyfulness to it.”

6. Learn from your peers

To help them navigate this fast-moving space, there’s a growing range of sources where leaders can learn from their peers. That includes JournalismAI, led by Polis – LSE’s journalism think-tank – which offers global case studies, a Fellowhip Programme and other training opportunities, as well as myriad of articles being written on this subject on a daily basis

For AP, the next step in their work with local newsrooms includes supporting five AI projects designed to “support of long-term business sustainability.” 

These efforts highlight some of the possibilities that this technology can unlock, including automated summaries of public meetings, translation of news alerts into other languages, as well as sifting through email tips and pitches to identify those that are the most promising and useful. 

The tools stemming from this work will all be made available as open source.  “Any newsroom can use it, any person could use it,” Rinehart says. “We have high hopes that from [these] project[s] could be a universal tool, or a tool that is meant to work universally… but it will work for more people than ever possible before,” she adds. 

To the AI-sceptics these efforts may help to further demonstrate the potential range of useful ways this technology can be used by users and content creators. 

Moving Forward

AI is not replacing journalism, or humans, any time soon. Indeed, recent developments highlight some of the challenges associated with this tech. 

Nevertheless, AI is here to stay and only going to become more significant as an instrument to help drive newsgathering, production, distribution and business operations. 

AP’s work in this space will continue to help lead these developments, both in terms of advancing their own business and supporting the wider media ecosystem. And as AI continues to evolve and become more significant in the media industry, the lessons that can be learned from these efforts will remain valuable for other publishers and media companies looking to harness the potential of this technology.