During the Sixth World Media Summit, the application of generative artificial intelligence and large language models became a hot topic among many media people. The media industry is ushering in a new stage of human-machine collaboration. Artificial intelligence (AI) is increasingly involved in the entire chain of news information collection, production and distribution. In this wave of intelligence, media professionals have also faced the new task of “upgrading” their technical literacy and knowledge structure.
The national high-end think tank research team of Xinhua News Agency released the report "The Responsibility and Mission of News Media in the Artificial Intelligence Era" during the summit. The report's survey of news media organizations in 53 countries and regions around the world shows that 10.2% of media organizations have fully embraced AI and have established corresponding mechanisms at the institutional level to introduce AI into the production process; 41% of media organizations are currently Actively explore the application of AI technology, and encourage and support some news business sectors to trial AI technology.
There is no doubt that newsrooms are becoming increasingly “technical.” When data journalism emerged, learning data processing and Python programming became popular; and now that the application of large models is in full swing, mastering the new language of communicating with AI models - "Prompt Engineering" has become a "required course" for media people.
"Prompt engineering" refers to the process of designing and constructing effective prompt information to guide the AI model to generate more accurate and useful answers or complete specific tasks. It is not easy to "train" large models. Journalists often need to design prompt questions just like carefully designing interview questions, so that the large model's answers can be more "to the point" and improve the accuracy of its functions such as finding news background information and conducting data analysis. Amedia, Norway's largest media organization, has begun spending a lot of time training employees on "prompt engineering", learning how to effectively ask large models questions, and developing work codes and training courses.
Reporters and editors are actively trying to use AI as a new partner to improve the efficiency of editing. The Xinhua News Agency think tank report shows that the top three application scenarios that the interviewed media that apply generative AI at the institutional level have explored or intend to explore are: first, auxiliary editing, such as fact checking, speech-to-text, translation, etc.; second, Create content, such as generating abstracts, making graphic posters, dubbing digital anchors, etc.; third, planning topic selection or drafting outlines.
Yan Lingsi, vice president of Reuters Asia Pacific, said during the media summit that Reuters is particularly excited about the potential of generative AI in three key areas: reducing mechanical work in the newsroom, using machines to enhance the work capabilities of journalists, and building and applying new technologies. AI tools to change the future of business.
Yan Lingsi said Reuters has integrated AI-generated headline assistants and bullet point summarization tools into its web publishing platform. “We found that the AI tool was very good at summarizing stories and generating headlines, and it was a very time-saving tool.”
Looking around the world's media, the application of AI has indeed brought improvements in the "efficiency", "quality" and "quantity" of news production. Xinhua News Agency's "Yuan Mao" metaverse system is driven by artificial intelligence to generate content, including multiple sets of production auxiliary tools such as Digital Man and Yuan Rubik's Cube; the news aggregation and content extraction system "Juicer" developed by the British Broadcasting Corporation uses AI to automatically capture The content of global free news websites is taken and classified to provide reporters with news materials and topic selection references; The New York Times developed a data analysis robot "Blossomblot" to analyze articles on social platforms and predict information suitable for dissemination on the platform to help create "hot hits" "Content; Illariy, an AI digital human jointly developed by the Peruvian Andes News Agency and the Media Laboratory of the Department of Literature of the National University of San Marcos of Peru, uses the local indigenous language for news broadcasts...
So, in the face of the "transformation and upgrading" brought by AI to the journalism industry, when more and more AI hosts and AI broadcasters appear, are media colleagues worried about being robbed of their "jobs" by AI?
According to Iqbal Sefer, chairman of South Africa's Independent Media Group, the relationship between AI and media content is like an architect and a building. The "architect" AI can only build the external framework of the "building". The internal details of the "building" cannot be completed by AI alone. It requires reporters to dig out the details, depth and humanity of the story.
Suresh Nanbas, editor-in-chief of The Hindu, holds the same view. In his view, generative AI cannot replace human creativity, but enhances it by providing new tools, expanding access, and unlocking new forms of content.
Pavel Negoitsa, president of Rossiya Gazeta, said that compared with humans, AI lacks some “personality.” “It is difficult to teach AI to possess the personal characteristics of each reporter, especially the individual characteristics of talented and distinctive reporters. Features".
At present, AI is still unable to possess the profound social and humanistic care, professional in-depth reporting capabilities and complex emotional understanding that journalists have. AI should "replace" trivial and repetitive tasks, allowing reporters to have more time and energy to engage in more creative and in-depth reporting. For example, many sports event broadcasts and financial market dynamics have been compiled and distributed by AI writing robots.
In fact, the deep integration of AI and news business processes is creating new positions, such as "AI news product leader", "prompt engineer", "fact checker", "AI audio and video editor", "AI model detection and tuning engineer", "forward-looking" Technical Researcher" etc.
The impact of AI on the media ecology is constantly evolving. For reporters and editors, instead of worrying about being replaced by AI, it is better to "dance together" as soon as possible, open their minds to embrace new technologies, and create more new possibilities for news reporting forms.