From audience analytics to programmatic advertising and automated story creation, media companies have used Artificial Intelligence (AI) for some time. However, this technology is rapidly maturing and opening up new creative and business possibilities that media executives need to be aware of.
In fact, AI is creating what venture capitalist and MIT fellow Paul Kedrosky describes as the “most disruptive change the U.S. economy has seen in 100 years.”
ChatGPT, an AI chatbot, is the current poster child for this robotic reckoning. Garnering a huge amount of column inches in recent weeks, the application can provide detailed answers to questions and prompts. Along with other AI-generated innovations like the portrait app Lensa and OpenArt – a gallery of works created by AI – these tools have inspired the latest wave of discussion about the implications of this technology.
Amidst copious innovation and optimism, concerns have also surfaced around AI-generated content, consent, bias, labeling and regulation, as well as the impact on labor markets. None of these issues are going to go away any time soon. Nevertheless, while media companies and policymakers navigate this unfolding landscape, the roll-out and adoption of AI continues to gather pace.
Artificial intelligence at work in the media
With AI having a real moment right now, this is the perfect time to explore the ramifications for media companies. Here are six uses of AI technologies that need to be on your radar:
1. Driving engagement
One of the most common ways publishers are using AI and machine learning is through AI-powered algorithms which personalize content recommendations.
This can help increase engagement and keep readers on your site for longer. That’s particularly useful if time on site is a key performance metric. Of course, it can also enable you to serve more adds to your audience too.
Personalized recommendation technology has long been the mainstay of platforms like Amazon, Spotify, and Netflix. Now it’s becoming increasingly common for other forms of content too.
One early proponent, The Washington Post, uses AI to personalize the news that they deliver based on readers interests and preferences. It’s an approach they’ve been using for some time across their app, newsletters and now the homepage.
As Digiday explains, the Post offers a personalized “For You” section on the homepage that taps into information provided during onboarding. At sign-up, subscribers or registered users can select their topic preferences. Recommendations are further augmented by your reading history and other performance data.
It’s an area the Post looks set to double down on, as they and other outlets seek to move to a more tailored content offering and away from the “one size fits all” approach of yesteryear.
2. Tailoring your paywall
One of these models, dynamic paywalls, deploys AI to change free article limits. As a result, users hit the paywall at different times, based on their behaviors and other indicators that help to determine a consumer’s propensity to pay.
There is “no magic number” after which readers will subscribe, notes Piano CEO Trevor Kaufman in an article that asked: “Has AI brought an end to the metered paywall?”
“Piano has seen visitors subscribe after a single pageview. Others take much longer to make the decision to convert, while some aren’t likely to ever subscribe at all,” Kaufman observes. In response to this variance, he argues, we need “smarter, more satisfying automation.”
AI can help. New York Media and Neue Zürcher Zeitung (NZZ, Switzerland) are just some of the publishers to adopt this model. They have used AI to determine individual paywalls, based on variables including geography, consumption habits and visit behavior, as well as subject matter and the device being used. Expect more publishers to follow suit.
3. Creating content
Many early newsroom experiments with AI focused on the potential to craft stories that typically follow a predictable formula.
One of the earliest to leverage AI for content creation, The Associated Press (AP) has been using AI since 2014 to generate summaries of earnings reports from publicly traded companies. This allows them to quickly and accurately provide readers with key information, freeing up reporters to do other work. “Prior to using AI, our editors and reporters spent countless resources on coverage that was important but repetitive,” their website notes, adding that this “distracted from higher-impact journalism.”
Alongside freeing up reporters, the technology has allowed AP to create more of this content. Automated story generation has enabled AP to increase the volume of these corporate stories by a factor of 10.
At a simpler level, AI is also being used to liberate resources otherwise hoovered up by resource-heavy work such as interview transcriptions.
AP is currently working with local newsrooms to help them increase their use of AI tools. In a survey asking what would be the most useful use of this technology, automating transcription came top.
4. Distributing content
A further potential benefit of AI can be seen in its ability to support publishers in their desire to get material in front of audiences – wherever they may be.
POLITICO Europe has used AI to convert two of their popular newsletters, Brussels Playbook and London Playbook into daily podcasts. The audio option gives subscribers another way to consume this content on the go.
This type of technological solution can help publishers manage their resources more efficiently, as well as distribute content to different platforms in a timely and cost-effective manner.
A further mainstream iteration of this idea is also being developed by Google. Dyani Najdi, Managing Director of Video and Display EMEA, has highlighted how the tech giant is experimenting with a tool to reformat landscape videos for YouTube. Viewers will see videos in square or vertical formats, with the shape automatically determined by how you are accessing the platform.
Although currently only available for certain video-ad products, it’s not a big leap to imagine this being used for other content in the near future. If it is, that would be a huge time-saver for many publishers. A further boon is the possibility of this technology opening up new distribution avenues, without the time and expense of repurposing everything.
Where we go from here: two trends to keep an eye on
The manner in which AI is being employed is constantly changing. Its possibilities have sparked discussion about the implications for education, journalism and other creative work, as well as the wider knowledge economy.
Within that, here are two key AI-trends for publishers to closely follow and potentially adopt.
1. Leveling-up content, and ad, personalization
Based on their interests and preferences, AI can personalize the news that publishers deliver to readers. Its usage is only likely to increase and become more ubiquitous.
More than 9,000 publishers use Taboola’s recommendation platform. Earlier in the year, they announced that AI functionality had been added to their homepage techstack. The company said that in beta testing companies such as McClatchy, The Independent and Estado de Minas in Brazil, had seen a 30% – 50% increase in clickthrough rates for homepage sections personalized by Taboola.
Alongside content, AI can also be used to deliver a better ad experience. Publishers like Condé Nast are using machine learning to find patterns that can lead to more personalized and contextual ads. In a cookie-less future this type of approach will be essential if ads are to be targeted and relevant.
2. Improving and streamlining workflows
With cuts being seen across the media landscape, a key challenge for publishers in 2023 will involve maintaining output levels (never mind launching new products and verticals) with fewer staff.
AI may help here, given its ability to be used for A/B headline testing and other forms of predictive analysis. It can also tag and generate content such as business, sports and real estate stories. Or, as seen at Forbes, provide detailed prompts for writers.
It can further support social media and off-platform strategies too. The South China Morning Post saved resources akin to work done by 3.9 full-time employees by using AI to streamline its social media management.
Meanwhile, in Germany, Frankfurter Allgemeine Zeitung has used AI to help editors understand which stories to put behind the paywall. This matters given their freemium model, and the need to balance free content that drives subscriptions with premium subscriber-only content that readers value.
The big picture
This list of uses is far from exhaustive. To it we can also add important developments such as the ability of AI to help address inequalities (through the automatic creation of audio articles, and work to measure gender disparity in news coverage), as well as the rise of automated fact checking and many others.
Although no one knows how this technology will play out, it’s clear that AI can play a valuable role in helping publishers with their operations. As a result, it is no surprise that key activities unlocked by this technology – such as data analytics and automation – are among the top investment areas for publishers in the coming year.
Previously, as the Knight Foundation has found, “when we talk[ed] about AI in newsrooms, we seem to lean heavily on the newsgathering part of the process and maybe do not pay as much attention to the product or the business side of the ecosystem.”
In 2023, that may begin to change, as we see an overdue shift in the thinking about the role that AI plays in supporting the strategic needs of publishers.
From shaping the content you see (Pink News’ positive news filter), to aiding with translations of new international editions (Le Monde’s digital English language product) and improving your SEO (Summari and other tools), AI is here to stay and increasingly integral to publisher strategies.
Against a challenging business backdrop, as outlets begin to focus more on areas like product, subscriptions and retention, AI’s contribution to a publisher’s success will become more prominent and important than ever.