Artificial intelligence is slowly but surely making its way into various spheres of human life. And if neural networks are perceived adequately in some areas of professional activity, then in others they cause heated debates about the ethics and correctness of using AI in work. This is what happens with journalism. If news content can be generated by neural networks, a reasonable question arises - why do we need journalists at all? However, not everything is so simple. In this article, we will consider how the use of neural networks in journalism affects the release of content and influences the work of journalists.
Can it be used
It is technically possible to use neural networks in journalism. Moreover, many publications are investing in initiatives to study AI technologies that will help journalists create high-quality, catchy content. But it is necessary to understand that AI will not be able to completely replace reporters, at least in the coming years.
Possibilities of using generative AI for the media:
-
automation of routine tasks – news generation, analytics, content plans, reports, work schedules, daily short publication schedules, newsletters;
-
creation of attractive and optimized headlines;
-
processing large amounts of data with highlighting the main points and structuring information;
-
correction of speech and grammatical errors;
-
analysis of the emotional coloring of the text for the target audience;
-
fast, high-quality translation of materials for international media.
It should be taken into account that neural networks cannot always guarantee the accuracy of facts, especially in complex analytical articles. In addition, according to journalistic ethics, you cannot trust an AI generator to write full author’s articles.
Application examples
Generative AI is used in almost all areas of journalism. A striking example is the German publication EXPRESS.de, where digital journalist Clara works in the news department. This is an artificial intelligence tool with a humanized avatar, with whom colleagues interact daily. Clara’s job includes structuring articles, conducting research, and summarizing information. In general, Clara took on the journalistic routine, leaving the creative part of the work to journalists and editors.
The Argentine newspaper Clarín also introduced an AI-based assistant to its staff, which they called UalterAI. The tool helps increase readers’ interest in the publication. It provides different formats for presenting materials for several target audiences, including a summary for generation Z. Statistics show that more than 30% of readers use the digital assistant’s functions. Interest in the news outlet after the transformation has more than doubled. Moreover, readers most often use a summary of publications.
The Chinese publication SCMP uses AI to analyze archives and create new materials based on historical data. Neural networks help journalists find connections between past and current events. Thus, for the newspaper’s centenary, articles were published based on archival materials collected over the years. Journalists would have had to spend many months structuring data from a huge amount of information. The neural network completed the job in a matter of minutes.
More examples of using AI tools for the media:
-
Generating news articles for writing texts about sporting events or financial reports, where the facts are known in advance.
-
Google Translate and other tools help translate texts for international publications.
-
Using AI to monitor social networks finds relevant materials and regional news from subscribers.
-
Editors use AI to proofread and edit texts, which improves the quality of materials.
-
Based on an analysis of the interests of the target audience, neural networks recommend topics and formats of materials that will be in demand by readers.
Another example of using AI in news publications is an issue of the Italian newspaper Il Foglio completely generated by a neural network. The goal of the experiment is to show the influence of artificial intelligence on humans.
What can be generated
AI expands the capabilities of editors and journalists. Despite the introduction of neural networks into journalism, wordsmiths should not yet fear that they will be replaced by robots.
What types of journalistic materials can a neural network generate:
-
short reports, summaries, analytical reviews based on the provided data;
-
product descriptions, press releases, content for websites, posts for social networks;
-
headlines and leads, introductions to articles;
-
images, infographics, collages, illustrations for articles;
-
writing scripts for podcasts, video reports, review videos;
-
summary of long articles with highlighting of theses;
-
transcription, automatic transcription with editing of recorded interviews.
However, when creating analytical reviews, interviews, author’s reports from the scene, the role of a person still remains the main one. Neural networks rather complement, but do not replace journalists. A robot will not rush to the other end of the country to be in the epicenter of events, highlight the main points and conduct an urgent report without preparation. Artificial intelligence will not be able to win over the interlocutor during an interview, feel his emotions, understand intonations, catch sarcasm. This is only available to professionals in journalism, and not to everyone.
Tools
According to the results of recent surveys, only 3% of the surveyed journalists have never used the help of neural networks. More than 50% of respondents use AI regularly, the rest use artificial intelligence in their work sometimes. There are many digital tools that work on the basis of artificial intelligence for content generation and editing.
ChatGPT is most often used in journalism to generate texts of various formats. It helps to write drafts of articles, press releases, news notes, posts for social networks. Journalists use the tool to generate introductions, create catchy headlines, and formulate leads. One of the main advantages is the ability to set the tone of the text and adjust the material to editorial requirements. ChatGPT is useful for structuring large amounts of information, combining disparate data into a coherent narrative.
In addition to photos from reports, generated images have increasingly been used in journalism to illustrate materials. This allows you to avoid disputes about the authorship of images and create unique illustrations that have not been published before. MidJourney is considered one of the best tools for generating images. With its help, you can generate unique conceptual scenes, metaphorical images, stylized portraits. This is especially valuable for author’s columns, analytical materials, essays. MidJourney adapts the illustration style to the topic of the article.
A useful tool for video journalism is Synthesia. This is a neural network that can generate videos with avatars voicing content with translation into different languages. Journalists use Synthesia to generate short news clips, presentations, video reviews without the need for presenters. This is especially convenient for local news. The neural network can voice the same material in different languages, adapt video content for an international audience. Synthesia speeds up the release of video stories.
For transcription, which is necessary for journalists to convert audio and video recordings into text, Otter.ai is popular. The tool automatically recognizes speech, distinguishes the voices of speakers, adds timestamps, which simplifies subsequent work with the material. Otter.ai is useful when transcribing interviews, press conferences, podcasts. This saves a lot of time compared to manual transcription. In advanced versions, the tool highlights key phrases, generates a summary, which helps the journalist quickly find important points in the text.
Using useful tools for journalists in Russia can be difficult due to restrictions on access to foreign neural networks for Russian users. This problem will be solved by the Chat AI service. On the site, with the help of chat bots in Telegram and VK, an integrated browser extension, you can use the most popular AI tools without restrictions. You don’t need a proxy or VPN to access AI generators.
Pros and cons of using
Journalists have access to powerful tools that can generate text, translate interviews, create graphics and videos. Neural networks speed up content creation and process large amounts of data. However, AI is not without its shortcomings and is not able to replace a “sharp pen”. It is important to understand in which cases neural networks really help, and when, on the contrary, they become a threat to the quality of materials.
Advantages:
-
increased speed of material release;
-
solving routine tasks;
-
automation of processes;
-
generation of ideas;
-
saving resources.
But neural networks are not without their shortcomings. First of all, they are not always accurate. AI can generate fake facts, make erroneous conclusions, not understand the context, and not know the local slang. It is important to remember that neural networks rely on existing data, so they can reproduce stereotypes, outdated information, and not know about new trends. In addition, texts written by AI often lose their individuality, sound formulaic, and lack the author’s style. Therefore, today, a neural network can be an assistant, but not a replacement for journalists.
Results
The use of neural networks in journalism opens up new prospects for editorial offices. AI speeds up data processing, processes tons of information, highlights key points, and frees specialists from routine work. However, personal expertise, critical thinking, creativity, journalistic talent, and authorial style remain the main ones in creating news content. It is important to find a balance between digital technologies and a human approach in order to preserve the essence of the journalism profession.
