AI and its Impact on the Future of Journalism

By Jeremy Gilbert, Knight chair in digital media strategy at Northwestern University's Medill School of Journalism, Media, Integrated Marketing Communications

I cannot think of another time in the last 40 years when journalism felt as important as it does now. There are countless challenges and opportunities facing society that journalists can and should be prepared to communicate with the general public, but in many cases, that will not be possible without artificial intelligence (AI).  

There is a critical need for better and more high-quality journalism, but for that to happen, the industry needs to be able to scale in a number of different directions that frankly are not possible without the adoption and incorporation of AI-powered automation. And to accomplish that, we need more people like students in Northwestern Engineering's Master of Science in Artificial Intelligence (MSAI) program who are willing to apply their technical skills in a field rife for transformation.

Automation and AI research can revolutionize journalism and how the public consumes media in a number of ways. Here are three areas where I think there can be the greatest impact.

1. Reporters need better tools

The world reporters work in today looks very different than it did a generation ago. Beyond the obvious technological advances, journalists are awash in data at a scale that never existed before. When you think about recent investigative journalism projects, you're often talking about people looking through hundreds of thousands of documents. One thing AI tools and techniques can do is help journalists look for patterns and find stories in these larger collections of data.

Today, natural language processing is used to find aberrations in data, and that can be the basis or a starting point for a story. What I anticipate moving forward is that reporters will have AI tools that will assess live trends. Instead of a journalist working to gather a lot of disparate information and sort through it, reporters will, eventually, be able to train a system to watch certain trends and notify them when something unusual happens.

2. News experiences need to be more personalized

Today, news organizations are able to learn much more about their readers than basic demographics. We can use AI to not only identify that a reader knows something about a particular subject based on their previous reading habits, but also use that knowledge to provide updates on that topic to that reader.  

For example, say you've been keeping up-to-date on the American Jobs Plan as it moves through Congress. We should be able to use AI to tell you what happened today related to the plan, and how it might be relevant to you and your community. Or, conversely, if you haven't been following the status of the plan, we should be able to tell you not only what happened today, but also provide the background and context needed to help you understand the significance. We can help create empathy around major issues by not abstractedly referring to groups that might be impacted by this legislation, but by concretely mentioning things you know about. That kind of personalization is only possible at scale through automation.

3. The structure and substance of stories needs to be reimagined

A reporter's job requires a combination of different things. They go out and find information, they ask questions, they observe, and they identify existing trends in data. They write a story or record one to audio or video, and once it's done they consider additional ways to distribute it, like with a tweet or post on Facebook. Moving forward, that's not going to be enough. Reporters need to start considering how to craft different forms of that story, and for that, they'll need automated assistance.

Take a smart speaker, for example. If someone asks their Amazon Alexa a question, they are going to expect an answer to their question, not a 1,000-word story that may include the answer somewhere within it. Stories themselves have to become answers to questions that news consumers are posing, and AI will need to help reporters understand what those questions might be. To take it one step further, those answers will also need to be tailored based on who it is that is asking the question. If a 35-year-old asks their smart speaker for information about George Floyd, the answer should be different than if a 9-year-old asks the same question. We’re moving toward a different paradigm of storytelling, and it's one that can only be achieved with AI.  

We are taking steps toward all three of these areas simultaneously. We're seeing new tools for improved data gathering, and larger news organizations have done some amazing work in creating formulas and then building dashboards to help reporters identify interesting stories. Those same organizations are also beginning to play with more personalization of content. When I was at the Washington Post, we were proud to make small strides in each of those efforts.  

The most complicated of the three will be reframing what a story looks like. As we get systems that understand in a more human-like way the kinds of news that we have, then it will become possible to ask some of the more interesting questions people might ask. Then we can tell stories that are both a traditional read and a virtual dialogue. This will take time, and it's important to remember that smart speakers are still in their infancy, and creating content for them is even more recent. The reality, though, is that we have a generation of potential news consumers growing up who will be expecting to control a lot of different things with their voice. That will force storytellers — and journalists in particular — to rethink what it is they offer and how they offer it to consumers.

These are interesting challenges facing not just the journalism industry, but society as a whole. There are a handful of data scientists and developers who went to major newsrooms in the past decade and did some incredible work, and I have every reason to believe that if we increase the number of talented AI-oriented developers and engineers in newsrooms, we will see even more improvements. The need is absolutely there and the technology is there, what we need now is more people who can use these technologies to make a real impact. For that I'm looking to the MSAI program.

What I love about MSAI is that the work students do is applied to the real world and across domains; it is not just theoretical or in the computer science discipline. MSAI prepares  its students to make a difference — the journalism industry is ready for their help.

 

Jeremy Gilbert is the Knight Chair in Digital Media Strategy at Northwestern University's Medill School of Journalism, Media, Integrated Marketing Communications. He previously was the Director of Strategic Initiatives at The Washington Post, where he led an award-winning team that transformed The Post into an international, digital-first news organization.