With the advent of the AI era, what kind of attitude should the news industry take to deal with it?

Mondo Technology Updated on 2024-01-31

To prevent getting lost, the elevator goes directly to the safety islandNewspaper man Liu YaEast A

On November 30, 2022, ChatGPT, a natural language processing tool developed by OpenAI, was born. On March 16, 2023, a press conference was held to demonstrate the function of its artificial intelligence model "Wenxin Yiyan". On August 31, 2023, the "Wenxin Yiyan APP" was opened to the whole society, and 12 hours after opening**, the "Wenxin Yiyan APP" soared to the top of the free list of the Apple Store, becoming the first Chinese AI native app to top the app store list.

As large language models, whether it is ChatGPT or the "Wenxin Yiyan APP", they themselves do not directly affect the dissemination of news. However, with the continuous development of artificial intelligence technology, more and more news organizations have begun to use natural language processing technologies such as Wenxin Yiyan to improve the way news is produced and disseminated, thus opening up a new track for the media industry.

1 Large language models will change the content production paradigm.

Recently, Lao Liu tried the "Wenxin Yiyan APP", and first saw that there was a module of "Newlywed Marriage Certificate" in the "Yiyan Treasure Box", so he tried to ask it to write a marriage testimony, which was quickly generated.

I have to say that this marriage testimony is still quite decent, with both ancient and modern references, as well as the integration of the professions of both parties, which is basically enough for witnesses who are not too demanding. (Of course, Lao Liu still wrote another article by himself, you can't use AI to bluff people!.))

Then, I asked the "Wenxin Yiyan APP" to write a news report about Jishi Mountain, Gansu. It was out in another 1 minute.

Although the format is not quite like a news report, all aspects are considered. At the beginning, the overall situation is summarized, and later, there are content sections required for common catastrophic reports such as emergency rescue, medical treatment, post-disaster reconstruction, psychological counseling, and disaster relief supervision, which can be used as the basic data collection and outline drafting for reporters to write news reports.

The advent of large language models has changed the traditional content production that takes a lot of time and energy to conceive, review, write, and edit. In the production of news content, large language models have broad application prospects.

First, large models can help journalists collect and analyze information more quickly and accurately, thus speeding up news production. For example, the above-mentioned report on Jishishan in Gansu Province generated by "Wenxin Yiyan", the latest figures of personnel and economic losses are "at your fingertips", which can help reporters save a lot of time.

Second, large models can generate diverse texts, making news reports more colorful. You can use it to write the first draft of a news story, edit and revise it, and finally present a different style of journalism to meet the various needs of your audience. Large models such as "Wenxin Yiyan" can also transform news reports into various forms of visual content, such as charts, animations, etc., which will help readers better understand and digest news content and improve the influence and attractiveness of news reports.

Needless to say,toLarge language modelsRepresented by artificial intelligence technology will beChanging the paradigm of content production,Triggering disruptive innovation in the media industry.

Taking "Wenxin Yiyan" as an example, it is suitable for multiple application scenarios such as literary and commercial copywriting, mathematical logic calculation, Chinese understanding, multimodal generation, etc., helping people to obtain information, knowledge and inspiration efficiently and conveniently, and can use its deep learning function to extract key information from a large number of texts in a short period of time, and automatically generate most of the content of news articles, improving the work efficiency of **.

In the history of communication, there have been many changes in the mode of production, from words to sound to sound. Today's AI model has brought new changes to the bottom layer of content production. We should actively promote the application of artificial intelligence technology in the field of journalism and communication to adapt to the new trend of the information age. For example, through the development of intelligent news production and dissemination platforms, the efficiency and quality of news production and dissemination will be improvedUse big data technology to analyze news information, understand the hot spots and trends of public concern, and provide data support for news production and disseminationArtificial intelligence technology is used to develop intelligent news interaction methods to provide users with more convenient and personalized news services.

2 The "metaphor of the cave" gives us implications.

In Plato's Republic, he uses the metaphor of the cave (Allegory of the c**e) to describe the basic imagination of human knowledge. Imagine having a group of prisoners in a crypt;They had been there since they were children, bound by chains, unable to turn their heads, only to look at the shadows on the cave wall in front of them. Behind them there was a fire and a path through the cave;There is a low wall along the path, which resembles a screen for a puppet show. There are certain people who carry various instruments through the path behind the wall, and the firelight projects the tools that pass through the wall onto the wall of the cave in front of the prisoners, and these instruments are models made from real objects. The prisoner naturally thinks that the shadow is the only real thing.

If one of them happens to be released and turns his head to see the light of fire and objects, he will feel dizzy at first, but it doesn't matter, he will slowly get used to it. At this point, he saw that there was a way to go, so he would gradually walk out of the cave and see the real world in the sun. It was only then that he realized that the world he had lived in before was nothing more than a cave, and what he had thought was real was nothing more than an image. At this point, he has two options, either to return to the cave or to stay in the real world. Eventually, for whatever reason, he chose to return to the cave in order to persuade his companions to get out of the cave as well. But his companions think he's talking gibberish and will kidnap him and maybe even kill him.

According to Plato, prisoners are like us, and they represent the state of human ignorance. The process of the prisoner coming out of the cave is likened to the process of attaining truth through education. Among them, steering is a crucial move. We can interpret the path of ascent and the contemplation of things above as the process by which the soul ascends to the knowable world and becomes a wise man.

For thousands of years, human beings have been diligently exploring the unknown, and their understanding of the natural world has been deepening. So,We have a delusion,Think as long as you are diligent and focused,In the end, you can always use the human oneReason comesExplain the worldBut it must be admitted that, at least until now, human beings have a very superficial understanding of the world we live in and the universe as a whole, and even little is known about human intelligence itself. The advent of artificial intelligence may change this phenomenon to some extent, and it may touch a world that is beyond the reach of human cognition. Or rather,Artificial intelligence may break through the limits of human reasonOnExtensionA new frontier in human cognition

3 Good human editors and reporters cannot be replaced by machines.

As soon as ChatGPT was born, it became popular all over the world. One of the hot topics discussed was which professions will be terminated by AI, and there is no shortage of pessimistic arguments. A year has passed, and domestic companies are scrambling to launch their own large models like a competition, but they have to face a real problem: we are very short of successful C-end applications, or large applications. Even, from time to time, social platforms are full of all kinds of joyful ridicule after people try out the big model - AI is talking nonsense again.

The principle behind the "Wenxin Yiyan app" is big data mining, analysis and deep learning. In terms of speed and accuracy, large language models clearly have an advantage over humans when extracting useful information from complex and boring data and processing it to a certain extent. At present, in terms of functional attributes, the "Wenxin Yiyan APP" has been able to better handle standardized and structured content.

But to date, large language models have not learned to think and ask their own questions like humans when it comes to personality attributes. The basic skills of journalists, such as questioning, questioning, and finding news clues in the other party's discourse, are not yet possible with existing artificial intelligence, including large language models.

Strictly speaking, large language models are information collation, not informationContent creationDue to the peculiarities of journalism, as well as the requirements of this profession for social skills, negotiation skills, empathy, responsibility, as well as creativity and aestheticsFor the foreseeable future, it is unlikely that those good human journalists who are competent will be replaced by machines.

A lot of the work of journalists cannot be replaced by large language models. The life of journalism is truth. Large models don't have on-the-ground coverage, scoop coverage, and therefore more valuable content. The temperature of the news shows the responsibility and feelings of reporters. If the large model lacks human thinking ability, human emotion, and deep understanding and empathy, it will not be able to write in-depth and warm news reports like a good reporter.

Lao Liu thinks,We should look at the actual ability of the large model dialectically in the present, and do not underestimate AiPotential to drive future growth.

It should be said that in almost all industries, including the media industry, most of the daily work is repetition and combination, rather than discovery and creation. It is for this reason that the emergence of large language models such as Wenxin Yiyan will inevitably reduce the demand for editors and reporters, resulting in the disappearance of some jobs. This will undoubtedly put pressure on some journalists, but it will also objectively force them to transform, strive to improve their own quality and literacy, and engage more in imaginative and creative work to adapt to the new situation.

4 Discriminating take-or-have is the right attitude.

At the beginning of the 19th century, in the midst of the Industrial Revolution, new machines were constantly being born in Britain, which took the jobs of traditional craftsmen and caused a large number of workers to lose their jobs and go bankrupt. The workers regarded the machines as the root of their poverty and smashed them everywhere in the name of the Luddites. The "Luddite movement" was not a reasonable way to deal with technological change.

Today, when AI has caught up with human reason and skill in many areas, we should learn to let go of stubborn biases and acknowledge and accept the result that human rationality is not the only, perhaps not the best, way to know or navigate reality. In fact, humans have used tools to do things they can't, and the history goes back at least thousands of years!At the end of the day, AI is built by and commanded by humans, except that AI can exert a certain degree of agency, rather than just being completely passive and obeying human commands.

At present, the technological breakthrough of AI large language models and their application and impact have set off a boom. Whether it is an editor or reporter who is engaged in journalism, or a college student who is studying journalism and communication, how lucky!In the era of the most intense change in the media industry, we should personally experience the reconstruction of the traditional information communication system, the continuous germination of new media formats, and the continuous evolution of the media ecology.

Journalists need to provide more inspiration and creativity, and large models can collect and analyze large amounts of data, but news workers need to make data-based macro decisions. In the era of AI, journalists need to carry out more creative mental work and become truly strong with the blessing of artificial intelligence.

However, Lu Xun's doctrine of bringing it should also be "calm, courageous, and discerning", and it should "use the brain marrow and let out the vision".

In areas involving human psychology, emotion, morality, etc., AI must function under human control. Although AI models can effectively improve the efficiency and quality of news reporting, they also pose challenges to news ethics and ethics. The new ethical issues involve journalistic ethics, professional norms, information authenticity, privacy protection, etc., which have a far-reaching impact on both the public and the public.

However, there is no need to rest on our laurels, any new technology will bring governance problems for a certain period of time, but this is not a reason for us to be afraid of technological development. Instead, we should govern in the process of development and solve problems in the course of development. The development of artificial intelligence is by no means a one-day achievement, and requires a long period of exploration and technological accumulation. A good product is used, and the more people who use it, the more it can help it progress.

Epilogue. Wenxin Yiyan and other natural language processing technologies are changing the way news is produced and disseminated, bringing both opportunities and challenges. We need to understand not only the advantages and characteristics of large language models, but also its limitations and shortcomings, so that we can make large language models better for us.

Since its release in March 2023, major companies have developed their own large models, lest they fall behind the times. The value of Wenxin Yiyan is to lower the threshold of AI, promote the inclusion of technology, and enable more people to take the train of AI and embrace the future.

In the coming 2024, I hope to see more cooperation with large models such as "Wenxin Yiyan", create more news communication applications and business models, and effectively promote the transformative development of the media industry.

(The author is a professor and dean of the School of Journalism and Communication, Nankai University).

Related Pages