Consider your mission and money, and avoid the hype

Dorian heads Teeming Media, a thought leadership communication consultancy focused on media technology, as well as The Verticals Collective group of media company founders and operators. He teaches media business at Columbia University’s J-school and the Zicklin School of Business at CUNY/Baruch.

With so much discussion around AI and media, how can you decide what (if any) of it to use in your news organization?

One way is to apply a business lens to filter out the hype. Ask yourself whether the given tool, technology, or partner will help achieve your goals — and whether the goal you’re achieving is likely to help the bottom line.

Here is some guidance, with real-world examples of companies using AI to positive effect.

Your mission comes first

AI is best implemented with a keen focus on how to use it to achieve your mission. First, determine whether any AI technology or tool will help.

Consider the finances

To achieve a mission, you need to pay the bills. Simply put, can the AI help you either save money or increase income (or cashflow)?

Increase engagement

A key way media companies improve their financial picture is by increasing engagement. More engagement leads to more subscriptions and renewals, more ads shown, and provides evidence for funders that you’re effectively reaching the desired audience.

Multiple news organizations use AI to help engage their communities. The McClatchy chain has automated real estate and high school sports reports they wouldn’t otherwise produce, which leads to more views on their newspapers’ sites.

And while some express fears that using AI summaries will cannibalize article reading, Aftonbladet’s deputy managing editor, Martin Schori, says their experience was the opposite.

“Almost half of everyone who saw the (AI) summary went on to read the article,” he said. “Those who did tended to read the articles deeper.”

Free staff for higher value work

AI can help free reporters and editors to do more of the high-value work that pleases readers, listeners and viewers.

Norway-based Schibsted, for example, uses AI to produce first drafts of articles about sports and celebrities in the hope of supporting “long-term business sustainability” for their local news organizations.

Other publications use AI for first rounds of proofreading, to create thumbnail images for newsletters, even to help create newsletters that are roundups of aggregated content. AI applications can streamline audio or video production by doing rough cuts. My company used AI to help review and summarize submissions for media awards cutting down on person-hours by an estimated 30%.

Schibsted says they’ve saved thousands of person-hours by using AI for transcription.

Consider building a chatbot

Chatbots carefully trained on a controlled set of content are a great way to give communities access to tailored information that’s easily accessed and consumed.

Affonbladet, which is based in Sweden, created an AI chatbot to answer questions ahead of EU elections based on official EU data and their own reporting. The initiative received a login rate ten times more than similar initiatives, according to Schor.

The San Francisco Chronicle provides subscribers with a restaurant recommendation chatbot based on their restaurant reviews — something they couldn’t do, otherwise (more on that below). A chatbot trained on podcast content can help increase its audience.

Travel industry publication Skift gives subscribers access to a chatbot trained on a dozen years of their archives. Co-founder and chief product officer Jason Clampet says the application increases engagement and provides his team info on their community’s interests, which in turn informs future coverage. And, he says, it costs “only a few hundred bucks” to field thousands of monthly questions.

Provide what you otherwise couldn’t

Schibsted also uses AI to help with data-driven investigations, such as one in their publication E24 that found Norway’s minister of science and higher education had plagiarized in an essay. 

“This revelation had not been possible without the help of AI, at least not to publish it as quickly as we did after getting the tip,” says Andreas Fosse, Head of Editorial AI at E24.

The AP has been developing AI applications such as summarized city council meetings and coverage of public safety incidents for its member organizations, many of which are strapped for reportorial staff.

Do a deal with AI companies?

Some major publishers have done multimillion dollar deals with OpenAI and others. Perplexity is offering publishers a cut of revenue from results that appear in their AI-plus-search engine.

You may not earn millions, but by allowing AI companies to ingest and train on your content, you could have a consistent revenue stream. Be careful with the contract terms, such as by specifying that the company can train their LLMs but not use your content for real-time results, and by inserting an expiration date for the relationship.

Or, you may wish to forego any such deal, as some publishers have, and include code to block AI crawlers (which may work — or not). Jessica Lessin, founder of The Information, says publishers “should be patient and refrain from licensing away their content for relative pennies.”

Decide what to paywall

If you have a paywall, how much of your AI-created or enhanced material should you put behind it? That depends on your goal. Skift, for example, allows three chatbot questions for free per month, then requires a subscription. 

You can place ads in the chatbot answers, and link answers to paywalled or ad-supported content. You can also require a login and capture some information from people you can use to provide offers and ads.

Caution: Put people in charge

Always make sure a person is overseeing every piece of what you create with AI. There are multiple examples of companies being de-listed from search for having too much AI-generated content, and it’s not hard to find examples of embarrassing mistakes.

I also recommend keeping your newsroom abreast of your use of AI, explaining the benefits and training them in how to use it well. I’m told that in many newsrooms there is a general fear or even hostility to using AI.  That reminds me of news organizations that ignored the early days of the digital revolution and today are out of business.

At the very least, trying AI now with an eye to using it well for business purposes while keeping the mission front and center is useful R&D. That doesn’t guarantee survival, but it should help with sustainability.

---

Cite this article

Benkoil, Dorian (2024, Dec. 10). To use AI, apply business thinkingng systems to keep journalists safe. Reynolds Journalism Institute. Retrieved from: https://rjionline.org/news/to-use-ai-apply-business-thinking/