Machine software will be used to write the news that we consume on our tablets and iPhones. That may sound like a ridiculous proposition right now, but it shouldn’t. According to a report by LSE’s Polis, some 71 news organisations across 32 different countries are already using Artificial Intelligence (AI) to assist them in their news reporting. The effects of this discovery are quite profound. They paint an encouraging picture for the future of an industry that is currently under siege from all sides.
To understand why the findings of the Polis report are so interesting, one must first understand just what AI is. Minsky and McCarthy the fathers of AI research defined AI as any task performed by a human programme or machine that if carried out by a human, would require the human to apply intelligence to the task.
News organisations such as the BBC, The Washington Post and Reuters are using AI softwares in many unique ways. For example the BBC uses an AI software called Juicer to extract data from a variety of sources through observing 850 global news outlets and assigning tags to the stores and organising them into four categories: organisations, locations, people and things. This makes it easier for a journalist to find the latest story on something like the coronavirus, through typing the name in and allowing Juicer to quickly search the web to provide a list of related content.
Additionally the use of AI tools to analyse a multitude of sources such as blogs, press releases, social media posts and images and video enables journalists to quickly get up to date on fast-breaking news developments and generate content that summarises changing situations, quickly. Ensuring that they are the ones breaking the story first, thus embedding themselves within the public conscious and potentially developing a positive reputation.
The Washington Post has been experimenting with automated news writing using Heliograf. Heliograf puts together news stories through analyzing data as it emerges. The information that is produced is then matched to relevant phrases within a story template and the machine adds the information to create a narrative which could be published across different platforms. This is similar to a process being used at Bloomberg which uses a tool called Cyborg for content creation and management. Such programmes can help reduce the grunt work for journalists and ensure that in a world that is driven by clicks and ad revenue, news organisations can more fairly compete with one another.
Of course, it’s not enough to just use AI to produce news, news organisations need to be able to combat the rise and spread of fake news. Thankfully, AI tools are being produced which enable content producers and publishers to identify fake news and reduce their impact on readership. AI systems are able to identify patterns of real data sources and real news content from those that have been artificially generated. The machine learning systems can serve as a first pass editorial system that can verify news items against additional sources, automatically providing verification against third party sources, ensuring that real news stories are reinforced and false ones are debunked. News aggregating software can then automatically put in truth-checking links and sources and rate news stories which are inbound with their likelihood of being true. AI and Machine Learning tools are also being used to delete/ weed out automated bots which propagate fake news by muting/deleting comments and flagging user generated content for moderation.
Of course, there are some things that AI cannot do. It cannot emit complex communication, it cannot give expert opinion, it cannot quickly adapt to a changing circumstance on site and it cannot get creative. These are all things that for better or for worse, human journalists can and continue to do. Be it through reporting live at a scene, negotiating with sources, pushing a source, interviewing someone, asking demanding questions of a head of state, and then putting all that together into an article that makes sense. These are the skills that are required to make a news story or an analysis piece really make sense. No machine can do that, not yet at least, and it cannot add the human element to the news through painting the picture of the person behind the story.
The evidence suggests that as time goes by and AI develops, it will play an increasingly important role in shaping the journalistic world. It can make the process of gathering information and breaking news stories, much more efficient through removing much of the grunt work from reporters themselves. It does not seem as though it will completely replace the human touch, but rather be incorporated into procedure to ensure that journalists work with the tools at hand to produce the best work that they can.