A visitor shakes hands with a robot at the World Robot Conference in Beijing in 2015. China has been hosting the event for three consecutive years since 2015 to put on display the bleeding-edge developments in robotics and engineering. (XINHUA)

---

 

When Google’s AlphaGo defeated Ke Jie, the world champion of the ancient Chinese board game Go last year, it took computer scientists by surprise. Because of the game’s complexity and nuance, they had thought it would take decades to develop artificial intelligence (AI) sophisticated enough to best a top Go player. Since its victory, AI has risen to the spotlight. In the media, AI writing software has generated heated discussion. Some even predicted that reporting will be one of the nine jobs in which robots will replace humans in the not-too-distant future. 

 
New production revolution
Through the practice of intensive—and at times risky and expensive—investigative journalism, traditional reporters attempt to uncover new facts and social trends. With their narrative talent, experience, creativity and intuition, they convert these facts into stories for their audiences.

As technology advances, the art of storytelling is also becoming a scientific endeavor that uses machine learning algorithms, drawing upon the vast field of linguistics and the study of natural language. AI algorithms are being composed that can convert facts into readable stories in a fraction of a second.

Writing generated by a computer program is formulaic and programmed with a standardized production process. Yet as it stands, the word “robot” makes the whole affair seem more complex than it actually is. The reality, as computer scientists say, is that automated journalism simply slots data into a template that has to be written by a human.

Nonetheless, robot journalists have an edge over human reporters in the following aspects: First, the standardization of news production increases labor productivity by saving human, material, and financial resources as well as time spent on traditional news production, including interviewing, writing, editing, typesetting, proofreading and publishing. Second, it could realize personalized information configuration to meet the needs of different users through smart labeling and aggregation matching in the vast pool of data to achieve accurate push.

Next, intelligent processing gives rise to news value because timeliness, accuracy, and geographical and psychological proximity of news are greatly enhanced. Finally, software could work 24 hours a day as long as the algorithm is in a normal state, which largely reduces the cost of news production.

 

Limitations of robot journalists
Human intelligence consists of “implicit knowledge,” such as purpose, perception, comprehension, imagination, inspiration, insight and aesthetics, and “explicit knowledge,” which features information acquisition and refinement as well as the formation and implementation of strategies. AI in reality is a simulation of the latter, so robot-generated stories lack in creativity and talent, skills that only humans can bring to the writing and editing process.

In the big data era, writing robots are employed to analyze and arrange the huge amount of existing data, then generate text according to a template. Considering how they operate, it is doubtful that they are capable of covering all news categories and themes. At present, robot journalists mainly cover sports events, financial reports, emergencies and other data-heavy stories. Such reports are often simple, stylized and regarded as low-level products in journalism. While writing programs may perform well at the relatively formulaic task of newswriting, it is hard to apply this technology to features and investigative reporting, which involve human emotion and complex situations, or opinion and commentary pieces that require value judgments.

In addition, in the process of news production, interviews and writing are inseparable for human reporters, while for robot journalists, interviews are absent, so the articles they produce often lack the sense of presence, inspiration, insight and value judgment. Meanwhile, due to their limited capacity to investigate and verify complicated facts, it is difficult for them to produce explanatory and critical reports.

More often than not, readers want to know the underlying background and details behind news events, and this part could only be accomplished by experienced human reporters. For example, if we want to cheer up a defeated sports team and let them know they are heroes in spite of their failure, we cannot pin our hopes on robot journalists.

Lastly, robot journalists cannot find a “human angle” to make a story interesting or filled with humanistic care. Though they may send out interview questions, they cannot detect the interviewee’s intention and subtle sentiment, so it is difficult for them to write a “warm” story. On the contrary, a journalist with strong news sensitivity is equipped to discover bittersweet and humanistic stories behind the scenes of trivial details.

As British computer scientist Simon Colton put it, if one wants to learn about facts, AI journalism is perfect. However, if a robot were to write a piece on the bliss and joy of childbirth, one may click the off button at the first few lines. Though the software has been designed to imitate various writers, they could never obtain the artistic temperament of those masters.

 

Role transformation
Human intelligence creates artificial intelligence for the purpose of attaining its services. The concerns that robots will steal human jobs are unnecessary. In a sense, work is not a necessity for human beings because the ultimate goal of work is to meet the material and spiritual needs of humanity. If robots could take over the mundane task of trawling through wire copy to spot a newsworthy item and other straight reportage of facts and figures, humans could get on with more investigative and analytical pieces.

In media, AI could promote the division of labor and synergy in news production, in which robot journalists will focus on the field of data journalism where they outstrip human intelligence while human reporters are dedicated to inspirational, creative, in-depth and warm stories.

For a long time, journalists have been labeled as recorders of news and facts, but the emergence of AI journalism has altered the landscape. In terms of recording, human reporters clearly do not have a winning game, given that speed and accuracy of robot writing are rather impressive, which means human reporters must transform their roles in news production.

For one, journalists should spend more time and energy on creative news, in-depth reporting, commentary and guidance of public opinion. For another, journalists should keep pace with the latest technology and familiarize themselves with computer software to achieve human-machine collaboration. Once robots complete analysis, process vast amounts of information and reach preliminary conclusions, journalists should step in to investigate and verify to play the role of participant, polisher, proofreader and the final gatekeeper of stories.

The rise of automated journalism software has undoubtedly provided a new direction for news production, which not only creates great opportunities for the transformation of traditional media but also expands the market for the development of network media. As time goes by, robot journalists will become more intelligent and the writing quality will grow closer to that of a human reporter. This will lead to the disappearance of some jobs in the newsroom, but new posts will come along, so journalists need to get ready for job reassignment and career changes.

In the process of role transformation, journalists should strengthen “Four Capacities,” as proposed by Huang Yuansheng, the first journalist in China in the modern sense, to make themselves irreplaceable in the face of automated journalism. First, “think with the brain” to make stories have a perspective and an angle. Second, “run with the legs” to bring stories closer to reality, closer to life and closer to the people. Third, “listen with the ears” to grasp the hinted message of interviewees and reflect subtle emotions in news writing. Finally, “write with the hands” to form a personal style and create a humane, in-depth writing based on a wealth of life and professional experience.

In sum, it is true that AI will change the way a newsroom works, and even bring about fundamental transformation to news production. However, no robot, at this point, can contextualize, piece together, and create original, evidence-based conclusions like humans can. Human reporters thus need to learn how to make better use of machines, so that machines become human tools. At the same time, journalists should always remember their professional ideals and carry a sense of humanistic care, which is also the driving force of the journalism industry.

 

Zhang Zhizhong and Ni Mei are from the School of Journalism and Communication at Southwest University of Political Science and Law.

(edited by YANG XUE)