With the development of artificial intelligence and the widespread use of robotic products, humans seem to be facing an unprecedented employment crisis. Although the situation is not obvious now, but after a few decades we may have been found in the depth of the dilemma.
Just as the appearance of the video erases the post of radio stars, artificial intelligence can also destroy writers, journalists and editors.
From Fiverr freelance to the "New York Times" reporter, a lot of text experts may soon find themselves unemployed. However, they are not defeated by overseas competitors, but by the algorithm to defeat. In order to understand the future of writing, and the appearance of artificial intelligence writers, we first need to understand those who have been replaced in the type of work.
Automation does not extend the reach of all jobs to the same extent, which is easy to see in American society. In the hinterland of Ohio, the post-industrial hinterland shows a very different picture compared to the highly concentrated intellectual capital city of San Francisco.
Despite the direct focus on overseas competition, but it seems that the largest work killer is not outsourcing, but automation. After all, the performance of the US manufacturing industry is quite good: manufacturing industry grew nearly 2.2% per year, far faster than the US overall economic growth. The US economy grew by 1.6% in 2016.
It seems that the manufacturing industry is booming, and the number of workers is not. But why is the work of the factory so susceptible to the impact of the robot? Why is it not a San Francisco-based programmer, or a New York writer?
This is due to the task. Organizations such as Oxford, McKinsey and PricewaterhouseCoopers have come to the conclusion that the most easy-to-automate jobs have several important things in common. They must have repeated routines and a high degree of predictability, the specific reference to the large warehouse assembly line or storage box. The job of these posts is simple and straightforward, and does not require adaptation or horizontal thinking.
On the contrary, with highly unpredictable work and complex problem-solving needs, it is unlikely to be accepted by the machine. The National Public Radio has a small tool that predicts the likelihood that your job will be replaced automatically. The score of this gadget to the writer is that 3.8% of the possibilities are replaced by a computer program. The traditional view is that creativity is not easily copied by the machine.
Is that right?
To make the work of artificial intelligence writers efficient, it must pass the Turing test, in this test, the computer must deceive the human, so that they mistakenly think it is also people.
This is especially important for creative algorithms where users do not want to use content created by robots because we believe that robots can not effectively communicate with us at the emotional level, and we believe that creativity does not exist in formulation, that is, we can not be simple To simplify the work of "war and peace" into arithmetic and binary input.
But the reality is that programmers can actually inspire creativity, and they have done it. As early as 2011, an undergraduate at Duke University modified an algorithm to break down poems into smaller parts (poems, lines, phrases) and then automatically generate poetry. One of them was even used by Duke University's literary journal TheArchive. Therefore, the work of the artificial intelligence writer was treated as a human work, and thus passed the Turing test.
Of course, in the "New York Times", a nine line of poetry and a lengthy article between the existence of a world of difference. However, it is important to realize that this is a key milestone. Over the years people have always thought that creativity is the machine can not touch. Nowadays, artificial intelligence has created poetry, songs, and even short films, which have become reality.
What will robot writers look like? Some of the views of artificial intelligence writers that the existence of unimaginable, one of the reasons may be that most artificial intelligence writers are difficult to achieve the similar level of human writers. For example, Facebook has shut down its artificial intelligence language building system because they have no way of effectively using natural language.
However, because of such events and ignore the artificial intelligence, it is very dangerous, because artificial intelligence has occupied a place. Not only artificial intelligence writers pass the Turing test, but they can also rely on specialized algorithms, such as depth learning, to hone their writing skills. The recent application of the depth of learning has made artificial intelligence in the Go contest to defeat the human. In addition, artificial intelligence has been able to deal with a lot of data seamlessly, and human flesh also need to eat and rest. For example, despite the initial setbacks, IBM's Watson has been able to analyze thousands of reports, generate their own views, and even help doctors do diagnosis, save lives.
This is a small step for the robot writer. In advertising, artificial intelligence copy is very common: they can draft hundreds of different advertising copywriting, test and analyze the advantages of each different iteration, the use of depth learning, and quickly create a good work. Most importantly, artificial intelligence does not need to rest, get paid, there is no prize ceremony that need to spend money.
In order to avoid repeating the same mistakes, it is clear that we need a new way of cooperating with people rather than competing.
In fact, the Washington Post has an article written by an artificial intelligence writer. Media executives turned to artificial intelligence writer Heliograf, to help expand their network audience. Editors enter Heliograf for keywords and templates for various news events. Subsequently, Heliograf searches for data on the web and matches the keywords, and then it generates reports, or reminds journalists to repeatedly check the errors in the text.
Heliograf can write simple reports of events such as elections or the Olympics. However, they can not do an in-depth analysis, which is the "Washington Post" after careful consideration of the decision. This article does not use some well-researched long-text narratives to attract small audiences. It uses Heliograf to create a large number of short mini-stories, increasing page views.
As an enhanced intelligence to work with humans, Heliograf is a more aggressive model of human-computer interaction. Humans still have the opportunity to continue research and write in depth articles such as reports of maternal mortality in the United States, as well as secret investigation of private prisons.
Even if the machine learning to make artificial intelligence and human writing ability to match, and can filter a large number of data, but the focus of attention and interviews with human content, but it is difficult to control. Reporters may be less likely to write some simple reports, the release of high-level analysis and investigation of the article.
However, the pain may be inevitable: the emergence of a new version of Heliograf may lead to large-scale layoffs, as the media may cut costs by laying off local journalists and sports journalists, even if the post of investigators is retained. Manufacturing may have some similarities: automation increases output, expensive manpower is fired, and the rest of the work requires more advanced qualifications or experience. Ultimately, the number of human welders is $ 25 per hour (plus benefits and leave), while the installation, maintenance and operating costs of the robot are only $ 8 per hour.
On the future of writing, one thing is absolutely clear: artificial intelligence writers have emerged. However, it is far from the danger that we can meet the layoffs and the corresponding pain caused by how much it is unclear. Indeed, people want to be able to interact with artificial intelligence to create an enhanced intelligence model, but that does not mean that writers can sit back and relax. After ten or twenty years, the writers may find themselves in the same dilemma as the factory workers who are now being replaced.
