Industry Insights

The AP’s Francesco Marconi on the State and Future of Augmented Journalism

By Kyle Basedow | Nov 21, 2017


Welcome to our Industry Insights interview series, where we pick the brains of the industry’s most influential thought leaders working to make an impact on the future of storytelling. For this month’s interview, we sat down with Francesco Marconi, the Strategy Manager and Artificial Intelligence Co-Lead for the Associated Press. We learned how the AP is leveraging various AI tools to augment journalism, and how humans and machines can work together to enhance reporting, research, writing and editing.

Could you give us a brief background about how you became involved in artificial intelligence, specifically augmented journalism?

In the summer of 2013, colleagues in the News department at The Associated Press made an interesting suggestion to the strategy team. They recommended working with a startup in the artificial intelligence space to automate the creation of certain news content.

A few months later, a deal was struck with Automated Insights to automate the production of narrative text stories directly from data, first in sports and, soon thereafter, for corporate earnings reports. That was the first of several AI projects deployed at AP, and marked my first exposure to the field.

Since then, the intersection of artificial intelligence and journalism became a key area of personal interest which led me to explore the practical implications of AI in the newsroom, but also join research teams at Columbia Journalism School and MIT Media Lab.

As one of the leading news agencies in the world, how is the AP helping contribute to the industry’s transition to AI? 

As a news cooperative, it’s crucial for AP to develop expertise in emerging areas such as AI and share that knowledge back with the industry at large. Our contribution is made in three key areas: industry training, experimentation and research.

We host webinars and bootcamps for our partners and members, where among other things we train them how to use AI powered tools such as Automated Insights for text automation, Wibbitz for automatic video creation and Newswhip for social media listening.

In terms of experimentation, AP spends a lot of time engaging with startups, universities and research centers to find creative applications of new technology. A recent example, is a collaboration with MIT Media Lab to measure public discourse by analyzing the level of attention the American President has given to certain issues on Twitter, as well as how other users have responded.

The third pillar of our knowledge-sharing efforts focuses on research. We launched AP Insights, the thought-leadership and innovation portal where we share best practices as well as journalism guides focused on emerging areas such as artificial intelligence, virtual reality and the internet of things.

In your experience, how have AI tools improved daily processes? Have they had an affect on the quality of journalism as a whole? 

AP has witnessed firsthand the impact of using artificial intelligence in our daily operations. It’s enabling us to increase volume of content but also to differentiate our reporting.

For example, we began using algorithms to produce automated earnings reports in 2014 which led to an increase of 12 times the financial news stories as before (from 300 to over 3,700), including for a lot of very small companies that never received much attention. Not only the number of stories went up, but also the error rate went down. What’s most impactful is that applying AI in this particular effort has freed up 20 percent of journalists’ time, allowing those reporters to engage in more complex and qualitative work.

In terms of differentiating reporting through AI, we are starting to use methodologies such as machine learning to mine data and find hidden patterns. AP can now surface new story ideas that otherwise could have been missed if we just used traditional investigative approaches. Recently journalists in the data team were able to analyze crime datasets and find the “most typical” incidents related to gun violence, including whether the shooting was accidental or not, if it involved a child or even if a police officer was involved.

When you are telling someone about augmented journalism for the first time, how do you describe it to them? How do they react/respond? 

Augmented Journalism refers to leveraging AI beyond automation of content and processes; it’s about expanding newsroom capabilities to create not only more volume but also more differentiated journalism. When journalists use those tools to enhance their reporting, research, writing and editing, we call it augmented journalism.

As is the case with any new, impactful technology, mass adoption can be preceded by mass confusion. Some of the key concerns raised during conversations about the topic are the risks inherent in unchecked algorithmic news generation, the potential for workflow disruptions and the growing gap in skill sets required to manage this new specialty area. These are important areas of consideration that we are actively studying and developing ways of addressing and mitigating those crucial challenges.

It’s a common fear that AI-powered machines will take people’s jobs. How can humans and AI work togother to improve journalism? 

The successful implementation of smart machines in the news setting is not about replacing humans. It’s about using new technology to create efficiencies and unleash human creativity.

AI could bring newsrooms to a new area, which will require reporters to learn new technical skills such as machine learning analysis, but also to be comfortable with new tools for text and video automation. In fact, roles like ‘Automation Editor’ and ‘Computational Journalist’ are emerging alongside an increased interest in the fields of data science.

Editorial leaders across the industry are already thinking about improved workflows that can enable faster journalism and create more responsive newsrooms.

Augmented journalism is currently in its infancy stage but is growing fast. Where do you see the future in the industry headed, and how can AI help adapt?

Streamlining workflows, taking out grunt work, crunching more data, digging out insights and generating additional outputs are just a few of the mega-wins that have resulted from putting smart machines to work in the service of AP’s journalism.

In the next five years, newsrooms will have an arsenal of AI-powered tools at their disposal, and journalists will be able to integrate smart machines into their everyday work. Machine intelligence will be able to do much more than churn out straightforward, automated news reports.

AI will allow reporters to analyze data; identify patterns and trends from multiple sources; see things that the naked eye can’t see; turn data and spoken words into text; text into audio and video; understand sentiment; analyze scenes for objects, faces, text, or colors—and more.

Want to know what the future of news consumption looks like? Download our full report!

Kyle Basedow