Bot-led journalism on the rise

Artificial Intelligence in the news media industry is being used in increasingly innovative ways, enabling quicker research, cross referencing of data and more.

Key Takeaway

The idea of robots writing the news has already been making headlines. But bots are beginning to take on a more significant role in some innovative publications as they work alongside human journalists to improve output and generate more efficient, accurate and smarter content.

AI in journalism is shaking up the media landscape, bringing about changes to the way stories are sourced, written and promoted – and offering opportunities for even greater and more precise coverage. Going forward, it is more likely that AI-driven bots will play a more significant role in the media industry, enabling the journalism industry to better predict which stories will do well and craft compelling keyword-driven copy.

News giants have in fact been using bots to augment the journalistic process since 2012. Algorithms enable automated narratives in newsrooms, save time and save money.

For example, respected publishers in the industry like the Associated Press use Wordsmith, a software developed by Automated Insights. Its natural language generation (NLG) platform turns data into insightful narratives and allows “analysts to transform data into role-based, written analytics. With powerful NLG technology and easy-to-use integration tools, Wordsmith delivers insights that are relative to individuals, organisation structures, and individual company’s overall goals.”

In the case of the AP, the primary objective was to increase productivity and efficiency in regards to generation of quarterly earnings reports from publicly registered companies in the United States. Every quarter, when public companies in the United States release their corporate earnings, reporters at the Associated Press need to plow through them in order to extract and report on relevant financial data. Previously, reporters were able to generate around 300 reports on average. This resulted in thousands of quarterly performance reports that don’t get coverage.  By incorporating the software into the newsroom, AP was able to produce 4400 quarterly financial stories.

Wordsmith transformed respective data on earnings from Zacks Investment Research into a clean published AP narrative story within seconds. In fact, the team at Automated Insights even configured the algorithm to generate narratives in the writing style adopted by the AP, giving it a more personalised and human touch. The articles hold the same quality and accuracy, if not more, when compared to AP standards. Aside from a note stating that an algorithm wrote the article, it’s quite hard to tell otherwise.

See how Automated Insights works:

Example report generated by Wordsmith:

Forbes and The Guardian employ AI to generate automated news

Other news giants like the Guardian and Forbes use an AI software provided by a technology company called Narrative Science, in which an algorithm generates automated news by feeding live data sets and content from older articles. Owing to the formulaic and data heavy nature of content in business news (stock figures and financial numbers), Narrative Science is able to self-generate the news stories.

The algorithms automatically transform data and visualisations into insightful stories—thus making insights easier to consume and act upon. Fed by external insights from thousands of articles and their impact on consumers, it allows companies to interact with their data, charts, and graphs, and in return receive real-time explanations, powered by advanced analytics, that can be easily shared.


AI technology powers the BBC

The BBC, like any news giant, is quite the storehouse for significant amounts of data, from daily news stories generated by the BBC, to feature stories, videos and large amounts of archival data.

In an attempt to bring all this data together and make it more accessible, the BBC uses a data extraction AI tool called Juicer.

The Juicer is a news aggregation and content extraction API. It takes articles from the BBC and other external news sources, automatically parses them with semantic tags, and further groups them into four categories: people, places, organisations and things (everything that doesn’t fall in the first three).

For instance, if a journalist is looking to find latest stores on the use of AI across sectors, or news stories on French president Emmanuel Macron, Juicer very quickly scans available data through BBC’s own data and public sources, and generates a list of related content. It can analyze massive amounts of data and find trends that humans would never have the capacity to identify.

The tagging is not manually or editorially controlled, but is an algorithmic process. The algorithm analyses the raw text of articles to find concepts according to the four categories, i.e. people, places and organisations that appear in the text. However, whether or not a term gets tagged is also dependent on the context in which it appears.

AI adds value to the media industry

Companies can take advantage of AI to monitor the conversations being produced by the press in the news and consumers on social media, and extract insights relevant to their brand based on keywords, using powerful AI-driven tools like Meltwater.

As we grow into a more networked society, AI software will continue to innovate the journalism industry and help writers stay on top of the news that’s most important to their audience. The purpose of AI in journalism is to increase accuracy in stories, hopefully to help combat fake news and to capitalise on limited resources by automating mundane and routine processes, providing a presence across multiple locations, and executing quicker data research. Robotic journalism can be seen as a healthy addition to humans in the media industry, by allowing companies to keep their costs low and margins high, and remain relevant by identifying trends before they’re on the rise.

Recent Articles