Using AI responsibly

The application of artificial intelligence (AI) in mainstream journalism needs to be positive, minimizing the negative aspects of AI.

More and more people - including journalists and media content creators - are using Generative AI (GenAI) tools to write news and articles. The increasing application of AI in life is a global trend and is recommended by experts, but it needs to be used responsibly and avoid abuse.

Wide application

After being popularized and commercialized globally since late November 2022, tools like OpenAI's ChatGPT have been deeply applied to every corner of life. In particular, the press has quickly exploited "super-intelligent" technology and "metadata" AI to serve their work - from writing articles, editing films to editing, publishing and newsroom operations.

Mr. HX - a journalist with nearly half a century of writing - is most impressed by AI's support in data processing. According to him, when he was working as a manual journalist, he had to search for books, newspapers, and documents, which took a lot of time and effort. Once, just to create an international feature page, he had to spend several days searching through the libraries of two major newspapers. Since the internet, finding data has become much more convenient and faster. Currently, finding data (which takes up half of the time to create an article) has become many times faster thanks to the ability to search, extract, and synthesize using generative AI.

Many world- famous scholars have their voices imitated by AI in an online propaganda campaign. Photo: THE INSIDER

Even photo editing departments for newspapers benefit from AI - such as background removal, background adjustment, lighting adjustment, especially removing unnecessary details in photos. At a discussion on AI applications in media, the deputy editor-in-chief of a major Vietnamese online newspaper commented: "AI is changing journalism in a more positive and humane direction". With the characteristics of Vietnam, the application of AI in mainstream journalism is taking place in a positive direction, minimizing the negative aspects of AI. It can be affirmed that no other applied technology is applied in life as quickly and widely as AI. The same is true for the journalism industry. To date, almost all major newspapers and news agencies in the world such as the New York Times, Financial Times, Bloomberg, Associated Press, ... have used AI in the news production process to save costs and increase productivity. The Digital News 2025 report by the Reuters Institute for the Study of Journalism (UK) based on a survey of 326 media leaders from 51 countries and territories shows that: 96% of newsrooms will continue to use AI for SEO, automatic translation, editing; 80% of newsrooms will use AI to improve and recommend personalized content; 77% will use it to create content and 73% will use it to collect news - including verifying information, developing data journalism and investigation...

Trend in 2025, 75% of newsrooms will invest heavily in voice development technology functions, converting articles in text form into audio files in many different languages ​​and tones; 70% of newsrooms will use AI to summarize content at the beginning of articles.

Potential risks

While news agencies have taken measures to limit the negative effects of AI, some social media sites are flooded with content generated by AI based on the subjective intentions of users. With its capabilities, generative AI also helps bad actors create content with complete media formats (news, articles, videos , etc.) that are false, fictional but seem real.

A typical example is the media coverage of "singer Taylor Swift and her football player boyfriend Travis Kelce unexpectedly appeared at the "No Kings Day" protest in Los Angeles - USA on June 14, 2025", accompanied by a picture showing the couple marching with the crowd. But later, analysts said it was a fake photo created by AI. During the conflict between Israel and Iran, photos of the wreckage of a modern Israeli fighter jet shot down by Iran appeared online, but were actually created by AI.

The danger is that with just an idea and money to buy a service package, anyone can create fake content - from news, articles to images, videos - created by AI on demand. Not only creating still images, combining photos, AI's video creation level also produces clips that could not be more real, even in contexts or with events that could not happen in real life. For example, the clip of a foreign leader sitting and eating "pipe intestines" on the sidewalk in Hanoi .

Research by Reuters warns that AI capabilities will increase the risk of fake news and bias in content production, especially when AI is involved in sensitive topics such as politics and society. In the countries surveyed, 36% of people are currently comfortable using news generated by humans with the help of AI, and only 19% are comfortable using news generated primarily by AI with human supervision. The UN's "Brave New World Report" stated: "AI is transforming the fundamental right to seek, impart and receive information, as well as journalism... It also brings risks. Generative AI enables the creation of misleading content such as deepfakes, undermining trust in democratic institutions." In May 2025, in an article titled "The Chaos and Credibility of Global Investigative Journalism (GIJN), the two authors Reed Richardson and Andrea Arzaba warned: "Today's smallest newsrooms can also take advantage of AI tools to gain powerful new reporting capabilities, but this technology also threatens the traditional news production model; at the same time, it gives bad actors a dangerous new weapon to spread misinformation and erode trust in the press."

In mid-December 2024, The Insider news site published an investigation showing that there was an online propaganda campaign using AI voice-fake tools to create "fake voice" speeches of world-famous scholars from universities such as Cambridge, Harvard, Princeton, the University of Bristol... "speaking up" calling on the US to lift sanctions on Russia and urging Ukraine to surrender to Russia.

Use only as an assistant

It can be said that the most reasonable, feasible and widely accepted solution for applying AI in journalism is to use it as an assistant, not to write content instead of humans. Reuters Institute emphasized: "Accurate, reliable information. That's what people want."

Source: https://nld.com.vn/su-dung-ai-co-trach-nhiem-19625062119051131.htm