The Rise Of Automated JournalismSeptember 11, 2019
Machine Learning (ML) and Artificial Intelligence (AI) are together in the process of revolutionizing the creation of journalistic content. An enormous challenge facing journalists in the modern world is the ever-expanding sea of data available for analysis.
The old business maxim, ‘what gets measured gets done’, could be due an overhaul. In a lot of cases these measurements need to undergo interpretation before they can be useful to business leaders. What gets reported can be acted upon. Here we’ll take a look at a few recent developments in the story of meeting these challenges of information distribution.
What The People Want
One of the hardest jobs for content creators can be finding topics that will engage their readership. Intelligent algorithms are now able to search for topics that will satisfy these requirements, making it easier to create articles that are likely to trend. This can save writers from wasting time researching topics that will gain little traction. A complimentary technology involves the automatic selection of relevant stock images based on the content of the article. This is useful for saving further time and resources from being wasted.
It is not only in the generation of content that these automated journalists are useful. The Washington Post found they were unable to keep up with the continuous retinue of unwanted comments posted on their articles. They now enlist “ModBot”, which is able to moderate user’s comments and automatically determine if they need to be removed. The systems effectiveness comes from how a user’s language choice is analysed, as well as the precedents set by actions taken upon similar posts by human moderators. Comments sections of news articles often become battlegrounds for some of the most controversial of opinions. This development addresses this problem and further allows companies to save valuable time and resources.
From Start To Finish
Perhaps the most significant example involves the ability for digital minds to actualize the publishing process from the initial research phase. This can increasingly be done independently of human intervention. Algorithms are now being trained to dig interesting content ideas out of the reams of available raw data. This unlocks the potential for otherwise-undiscovered insights to become known. Using in depth statistical analysis techniques, alongside parameters for determining newsworthiness, these technologies represent a milestone in the history of content automation.
After the data is gathered, the analysis is performed and any ‘outliers’ are flagged. This is then paired with either the automatic or manual setting of topic importance, which specifies which of these articles will form the subject of the content. The system then crafts sentences based around the data, and publishes once it has confirmation from a human ‘supervisor’. With this, intelligent document-scraping and data-analysis come together with natural language generation to produce truly computer-led content.
An Automated Future
One notable humanitarian application of this technology was spearheaded by the LA Times. They are now able to report on earthquakes within minutes of the occurrence. Whilst not perhaps the most profitable application for the technology imaginable, it shows how the automated gathering, compiling, writing and sharing of information has the potential to transform lives in a myriad of powerful ways.