Introduction
In the digital age, automated technologies have disrupted the traditional writing standards and information gathering processes of journalism. Mainstream news sites have slowly implemented these programs into their online publications. While automation can certainly increase a writer's productivity, artificial intelligence has the potential to fully overtake the roles of reporters and editors. The “robot reporter”, as coined by several news outlets, is an emerging service in the journalism industry. AI is being used by The Washington Post, Reuters, Associated Press, The Guardian, and Bloomberg among others. In 2016, WaPo debuted an AI-powered storyteller named Heliograf, which wrote over 850 articles in one year. While that is considerably more publications than required of a typical journalist, Heliograf hasn't reached the more "human" side of reporting, like conducting interviews or in-person research. Investigative reporters, columnists, and lifestyle writers seem to be safe because they require some level of human-to-human interaction. Even more limited are the programs used by Reuters and Bloomberg (Bertie and Cyborg, respectively); their sole focus are basic financial reports. Extracting numbers from the stock markets and scores from baseball games are not the most difficult thing to compute. Translating that data into traditional writing formats without human intervention is a landmark feat, and a blow to journalists. Hard news may be the next field to be taken over by AI technologies.
Technical Features
Heliograf and related programs analyze, organize, translate, and visualize data in a comprehensible, ready-friendly format. Typically, an algorithm examines the given datasets, selects a pre-programmed article format, organizes essential points, and inserts information like names, locations, quantities, and statistical figures. The output may be tailored to a specific style or voice. There are currently seven AI models used in news media: AI fields, NLP, expert systems, speech to text, robotics, vision, planning/scheduling/organizing, and machine learning. Often they are used in conjunction with one another if the task requires broad capabilities. AI fields and NLP are used for data extraction, classification, and translation. Robotics and vision are employed in image recognition and machine learning. For predtive analytics, a tool would employ both the machine learning and AI fields models (1). Most AI tools used in journalism are structured with the NLP model, which combines Natural Language Generation (NLG) and Natural Language Understanding (NLU). This model gives the tool immense capabilities; they can extract meaning and significance from human speech and written text. From there, an NLP tool can translate, classify, and cluster the information.
Necessary Infrastructure
An artificial intelligence system is required to automatically assemble a news report. The AI software can be programmed to sift through heaps of data, thereafter performing quantitative analysis. The system should include a series of parameters in order to keep the tone and style of the article in conjunction with that of a standard news article. For such task, one or more article templates should be created and adopted into the software before performing any operation.
Workflow and Personnel
The procedures used in today's AI writing tools are a bit more complicated than the Tin Man churning out papers from a typewriter. It indeed requires human effort at initial stages. Writers and editors must methodically develop numerous versions of a news story, replete with text for potential beginnings and endings. The AI system can assemble an article once the data flows in. Every hard news article follows a standard format: a lead that summarizes the events in one or two sentences, the five W’s (who, what, when, where, and why), less important information, and a quote to conclude. This “inverted pyramid” of storytelling can be constructed with AI by inputting useful points of information. Axios, an American news site founded in 2016, breaks from the traditional storytelling fashion by using bullet points to summarize key elements of a news story.
Advantages
As technology improves, a common sentiment shared by workers is the fear of being replaced. If AI models gain advanced writing capabilities, this may be the direction of journalism in the coming decade. AI may develop more straightforward methods to write articles about a day's events or breaking news. As media conglomerates dance with the idea of replacing writers with algorithms, freelancers should certainly be cautious if AI technologies improve soon. According to surveys, the industry seems split about computer-generated writing. 43% are not at all worried or not very worried, 52% are somewhat worried or worried enough, and 6% are very worried.
Reference
https://www.researchgate.net/publication/355044842_Artificial_Intelligence_in_News_Media_Current_Perceptions_and_Future_Outlook