When seeing Tencent Finance's robot manuscript announcement, the first reaction of many editors was surprise, followed by secret worries about the future of their careers. When machines begin to enter the field of text work, will it bring about efficiency changes or a collision with human creativity?
Robot writing is nothing new
Robot writing is already widely used in overseas media. The Associated Press of the United States has been cooperating with Automated Insights since 2014, using its AI platform to automatically generate corporate financial reports. This cooperation is not experimental, but a mature daily production activity. The Associated Press is able to produce about 3,000 company financial reports every quarter, far exceeding the 300 that were previously written manually. This not only greatly increases production, but also ensures the accuracy and speed of data processing. Yahoo uses this technology to report sports events. It can automatically generate football match reports and update scores. In this way, it can quickly meet readers' needs for real-time information.
How technology generates a story
"Natural language generation" technology is the core of robot writing. This technology first inputs structured data into the system, such as revenue data, profit data, earnings per share and other figures in financial statements, as well as preset report templates and grammatical rules. Then AI analyzes the logical relationship between the data, such as whether a certain data is increasing or decreasing year-on-year, and then converts the numbers into smooth sentences and paragraphs according to the rules. The entire process is highly automated. It only takes seconds from the time the data is accessed to the final release, and it can also ensure that there are no calculation errors and the format remains consistent.
The value of human editors has not disappeared
Although robots are good at processing standardized information, in-depth analysis, on-site interviews and feature writing are still the exclusive domain of humans. After the Associated Press started using robots, human labor was not replaced but shifted to more challenging work. For example, they will screen out newsworthy companies from machine-generated financial reports and conduct in-depth interpretations or follow-up reports. Robots free up editors' time to process boring data, allowing them to focus on investigative reporting, exclusive interviews, and other content that requires critical thinking and emotional resonance.
Exploration and limitations of domestic applications
In August 2015, Tencent Finance released the robot article on CPI, which was a high-profile attempt by domestic media in this field. It marked the official entry of automated writing technology into the Chinese content production scene. However, current domestic applications are mostly concentrated in limited areas such as sports event news and simple financial data broadcasts. Compared with overseas, Chinese grammar is more complex and the context is more changeable. This brings greater challenges to the semantic understanding and expression fluency of robot writing, and limits the rapid expansion of its application scope.
The dual advantages of speed and accuracy
Speed and accuracy are the most prominent advantages of robot writing. In a public test in 2015, Automated Insights' AI and a human reporter simultaneously wrote text messages based on Denny's restaurant company's financial reports. The robot only took two minutes to complete a manuscript with accurate data and smooth sentences. However, human reporters need to spend more time checking data and organizing text. When dealing with massive and high-frequency data publishing tasks, this efficiency advantage is unmatched by human power.
Opportunities and challenges for future development
In the future, robot writing is expected to be applied in more vertical fields, such as generating regular weather reports, real estate listings, or simple product instructions. Its development also faces a clear boundary, that is, it is difficult to understand complex human emotions, cannot carry out creative ideas, and cannot write comments with a unique perspective. The direction of technological progress should be to better serve as assistants to humans rather than replace them. How to make AI learn to process more subtle contexts and perspectives while ensuring efficiency is a problem that developers need to continue to overcome.
When you read a fast and accurate financial report, do you care that its author is not human? In an era of information explosion, do you value more the immediacy and accuracy of content, or the depth and warmth of content? Welcome to share your thoughts in the comment area. If you find this article inspiring, please like it to support it.
