Peek in to the Future of Artificial Intelligence in Journalism

Whether it’s a retailer’s mobile app providing shopping recommendations based on past purchases or a self-driving car, artificial intelligence (AI) has become ubiquitous in modern society. And this newfound reliance on the scalability, accuracy, and personalization AI enables is only projected to grow — TechJury reports the global AI market is expected to reach $60 billion by 2025.

For journalism, a field that relies on the swift dissemination of accurate information, AI and machine learning (ML) present valuable opportunities to make reporting more efficient. In response to this potential, the London School of Economics’ journalism think-tank, Polis, has begun a project called JournalismAI — “a global initiative that aims to inform media organizations about the potential offered by AI-powered technologies.” In its AI Starter Pack, JournalismAI attempts to answer the most frequently asked questions about AI and ML in journalism.

What is artificial intelligence?

JournalismAI defines artificial intelligence as “an umbrella term to refer to the use of algorithms and automation by news organizations, usually to make journalists’ work more efficient or to deliver more relevant content to audiences.” More generally, AI refers to intelligent machines that can, at least theoretically, “think like humans.”

What is machine learning?

JournalismAI classifies machine learning as a subset of AI developed to process data and learn patterns. Powered by this information, ML-trained systems can perform tasks and answer questions without specific instruction from a programmer.

AI in the Newsroom

Initially, the newsroom may be the last place you would expect to see AI or ML. However, it’s already in play at media organizations globally. To better understand how AI is currently being used in news media, one of journalism’s leading nonprofits, The Knight Foundation, recently analyzed 130 AI-powered projects. Here’s what they found:

• 47% of the surveyed projects used AI to augment reporting capacity. From social media to document dumps, it’s impossible for reporters to keep up with every possible source for their next story. Augmented reporting enlists AI and machine learning to process large volumes of data and identify potential news stories.
• 27% used AI to reduce variable costs in journalism. AI-powered tools that automate tedious tasks like tagging and transcription save valuable time — and money.
• 12% used AI to optimize revenue streams. For publications still feeling the loss of print media, the extra revenue AI can generate through “dynamic paywalls, recommendation engines and the digitization of a news organization’s archives” is an significant boon.

The Knight Foundation also looked at the places AI was most commonly implemented in the news pipeline. They found 67% used AI for newsgathering, automatic story generation, and news production, but only 12% leveraged AI for product development, subscriber management, and paywall optimization. This indicates an opportunity for further automation and efficiency in the industry — “when we talk about AI in newsrooms,” says the Foundation, “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.”

AI Use Case: The Financial Times

To better understand the impact of AI, Forbes compiled a list of the most popular applications of AI in journalism in 2020. The Financial Times (FT) was referenced multiple times as an example of successful AI implementation, so let’s take a closer look at the different ways they put this burgeoning technology to use.

Trend spotting

When it comes to cutting through the noise to find engaging, topical stories, AI can’t be beat. FT specifically tailors its algorithms to spot market trends, in turn informing content creation and ensuring the publication keeps its finger on the pulse of the economy.

Reducing bias

For any reputable publication, journalistic bias is at the top of the mind — but it can be a difficult thing to quantify. That’s why FT implemented Janetbot (named after Janet Yellen, former chair of the U.S. Federal Reserve) to monitor the ratio of male to female faces appearing in the publication.

Proofing

There’s no room for inaccuracy in journalism, and AI is a great defense against the dreaded error. FT not only uses AI to “spot and correct errors,” the algorithm also tracks reader engagement and feedback.

Subscriptions

FT began charging for online content back in 2001, but in recent years switched to a subscription model — AdWeek reports digital subscriptions to FT increased 6% year-over-year in 2020. AI not only optimizes the subscription model by delivering a more personalized user experience, it can also help create more of the content subscribers want to see.

The Future of AI

As we’ve learned, the current uses of AI in journalism run the gamut from automating transcriptions to identifying trending topics — but what about the actual writing? It might seem like sacrilege to have a machine write an article, but so-called “robo-journalism” is on the rise, and has its merits.

Take What’s New In Publishing’s example of Swedish news publisher MittMedia. When Head of Content Development Li L’Estrade discovered real-estate articles were performing especially well, she set out to publish more content about properties. But with the sheer volume of houses being bought and sold, it wasn’t sustainable for MittMedia reporters to produce content on every listing. So they developed Homeowners Bot, an AI system capable of analyzing properties and writing short descriptions of them. As a result of implementing Homeowners Bot, MittMedia generates 480 articles on home sales per week and has converted almost 1,000 paying subscribers.

Reuters suspects robo-journalism will continue to gain traction as the technology becomes more fluent. “Every year sees more spectacular progress in the world of Natural Language Processing and Generation. In 2020 OpenAI came up with its GPT-3 model, which learns from existing text and can automatically provide different ways of finishing a sentence (think predictive text but for long-form articles). Now Deep Mind, which is owned by Google, has come up with an even larger and more powerful model and these probabilistic approaches are making an impact in the real world.”

But journalists don’t have to worry about robots taking their jobs anytime soon. The main benefit of robo-journalism is that it takes care of busywork, allowing journalists to focus their time on more thought-provoking pieces. Lisa Gibbs, Director of News Partnerships at The Associated Press sums it up best: “The work of journalism is creative, it’s about curiosity, it’s about storytelling, it’s about digging and holding governments accountable, it’s critical thinking, it’s judgment — and that is where we want our journalists spending their energy.”

Investment research publishing is another area where ‘embedded intelligence’ is playing an increasingly important role. Find out more.