At the moment of information explosion, the reading scenes of college students are undergoing disruptive changes - with the help of AI tools, they can not only quickly expand their general reading knowledge, but also easily complete in-depth thinking homework in class, and learn the general content without having to open the pages.

  In study and life, how can college students use AI tools to read better? Will AI-assisted reading lead to thinking dependence? At a time when technology is changing with each passing day, how should college students read it? April 23 is World Book Day. A reporter from Science and Technology Daily talks with college teachers and students to explore how to find a balance between technological dividends and reading thinking.

　　A capable assistant

  With the rapid development of AI, Liu Zhiyuan, associate professor of the Department of Computer Science at Tsinghua University, produces hundreds of papers every day in his field. "This forces us to learn to obtain the content we really need in the complex information. Mastering new reading methods, being good at using tools, and learning to distinguish the authenticity has become a compulsory course," said Liu Zhiyuan.

  Against this background, AI-assisted reading provides college students with many conveniences and innovative experiences. Yang Ziyue, a graduate student at the School of Education of Peking University, told reporters that when reading papers, he usually pays more attention to the core content of the experimental part. If it is too time-consuming to read the paper in full and then perform classification and summary, with the help of AI tools, it can quickly extract key information or lock in target content, and then expand intensive reading.

  Compared to what specific content has been read, Chen Ben, another graduate student in the school, believes that AI can better improve the efficiency of data integration. "When reading new information, AI can help me efficiently capture and integrate a large amount of related content, which is both time-saving and comprehensive," said Chen Ben.

  The reporter observed that AI-assisted reading tools such as KIMI and Yujing have become reading partners for many students. Students have different levels of use: some rely entirely on the content output by AI and do not read the original text; some use AI to analyze text fragments, and at the same time compare the overview of AI generation with the original text to read; some ask AI questions to fragment words and sentences, and then return to the original text to read based on the explanation.

  Chen Ben said that rapid reading and digesting a large number of strange texts can help understand a new field in a short time. The outline, summary or keyword generated by AI is like a preliminary "navigation map". "This 'map' may not be perfect enough, and may even have fallacies, but it at least provides an entry point, an 'anchored point' to start reading, so that students will not feel at a loss when facing the vast amount of information," said Chen Ben.

  "AI-assisted reading tools can be understood as search engines combined with iterative upgrades from AI. We should explore new opportunities brought by this technology to learning and knowledge acquisition." Liu Zhiyuan said, "AI's powerful extraction and correlation capabilities can help students get rid of explosive information and focus on classic reading. In addition, AI can also transform different languages ​​and obscure words into plain expressions, lowering the threshold for students to read in a general way."

　　Shortcuts to be lazy

  However, while AI-assisted reading brings convenience, it also raises many worrying problems. "Easy to become lazy" and "losing criticality" are high-frequency words that reporters hear in interviews.

  Yang Yuxi, an undergraduate student in the Chinese Department of Peking University, said that liberal arts students usually have to face a lot of reading tasks. After the big model appears, being lazy becomes a breeze. When it is difficult to complete reading assignments by "reading in person" but faces the academic requirements of class discussion and submitting reading reports, reading content processed by AI has become a very tempting and convenient channel.

  The reporter found that with the help of AI tools, it is easy for students to skip active thinking and deep processing, and overcome all the carefully set reading levels that the teacher has set in almost no effort. In response, Zheng Lei, assistant professor at the School of Education of Peking University, said: "The learning behavior of using technology to take shortcuts such as pieced together reading reports and online plagiarism have always existed, and AI has only amplified this old problem."

  Tan Ziwei, who studied in the Department of Philosophy at Peking University undergrad, said that in the process of trying AI-assisted reading, the most important thing to be wary of is to blindly believe in the content provided by AI without criticism. "When I first started to learn advanced modal logic courses, I thought AI tools were great. But as I learned, I found that AI often made mistakes in complex derivation and case analysis. If you blindly believe it, you can easily be misled."

  Zheng Lei believes: "AI feedback is mirror-based. If you are stronger, it will be stronger. If you are weak, it will be weaker. For beginners, the more they rely on AI, the more likely they will have an extremely one-sided or even wrong understanding of the text. Using AI to read without limitations may increase the gap in people's reading ability. For people with professional judgment, they will constantly raise challenges in the process of interacting with AI, which will inspire some interesting thinking."

  "AI-assisted reading does not mean that students do not need to read or think. On the contrary, it puts higher requirements on students." Liu Zhiyuan emphasized that college teachers and students need to realize that as an emerging thing, AI has problems such as flattery, hallucinations and database pollution that exist need to be continuously optimized. Even if AI technology is mature, students should effectively correlate the information produced by AI and make their own thinking and judgments, and cannot give up subjective initiative.

　　Is it a "person" or a tool

  Whether to use AI to assist reading is an active choice based on the purpose of reading. "We will not let AI help watch cool dramas. The content that we are really interested in reading will not be replaced by machines. The same is true for academic reading." said Lin Yichen, who studied in the History Department of Peking University undergrad. Lin Yadong, a master's student at the school's School of Education, also felt the same way. He believes that reading and reading are two different concepts. Reading is a spiritual experience that is driven by the feelings of "I" and thinking. When you care about personal feelings, you generally don't use AI; while reading is more like a hardworking job of extracting information but not processing. Traditional reading is like a farmer digging the ground with his hands, and using AI to assist reading is like having a useful hoe, which can improve efficiency.

　　So, in the AI ​​era, how should college students read it?

  The role of "leader" is crucial. The interviewed students generally mentioned that although AI is rich in knowledge, its understanding of the logic of text is completely different from the learning experience as a "person". Yang Yuxi was deeply touched by this: "When asking the teacher for advice, the teacher's past training experience, experience and methods, the teacher's eyes, tone, and random questions will all have an impact on me. In addition, some content may be rewarded for the first time, and you can only understand it after repeated thinking. How to understand the lag in this reading also requires communication and collision with experienced teachers."

  When talking about how to be a good "guidor", Zheng Lei said that the AI ​​era is increasingly amplifying personalized learning and guidance, but reading text is also a medium that connects oneself and others' experiences. Teachers should respect students' interest in reading and encourage them to form reading groups with like-minded friends, read and discuss together, and establish a real connection. In addition, when the purpose of professional text reading training, teachers should consider that understanding the context of subject research and grasping the core issues of subject fields is threshold and gradient. They should carefully design the "scaffolding" of reading tasks, guide students to use AI reasonably, and help students overcome the boring, incomprehensible reading, and even want to give up.

  "In the future, AI will not only be a tool for assisting reading, but will also become a reading partner that can engage in ideological dialogue." In Liu Zhiyuan's view, with the upgrading of technology, AI assisted reading does not conflict with traditional reading and reading clubs, but should collaborate around cultivating students' core abilities such as induction, critical thinking, and construction of views.

  "Who to read with and discuss with us sometimes has a greater impact than "what to read and how to read"." Zheng Lei said. (Intern reporter Jing Xiaoqing)

[Editor in charge: Zhao Yang]