Liu Haoran, a student at Northwest Normal University, believes that algorithms are everywhere. "Sometimes, just chatting with classmates about a certain topic and opening your phone, you will find that relevant recommendations have already appeared in various software. Liu Haoran is not very satisfied with such recommendations, "feeling that privacy is exposed in front of algorithms. Recently, China Youth Daily and China Youth School Media conducted a questionnaire survey on the topic of algorithm recommendation for young people, mainly college students. 3694 valid questionnaires were collected, and 78% of the respondents believed that they needed to reduce their dependence on algorithm recommendation. 72.8% of the respondents had encountered the situation of "big data killing familiarity", and 68.9% of the respondents believed that recommendation algorithms should be more transparent. Not long ago, when "algorithm recommendation" was no longer "invisible", Dou Yinghui, a student at Beijing Forestry University, talked to her friends about wanting to buy a pair of headphones. Her social media and shopping app pages were immediately occupied by various headphones. Sometimes Dou Yinghui would joke with her friends and say, "Speak softly, don't let my big data hear you." Ni Mengna, who just graduated from a university in Fujian this year, found that algorithm recommendations have permeated every corner of her life. "The videos recommended to me always make me feel comfortable and happy; The ride hailing app knows her daily routine very well, even knowing that she has to go to a friend's house every Wednesday or Thursday, and on that day, it will automatically fill in the friend's address in the destination bar. According to the survey, 89.7% of respondents are aware of the existence of "algorithm recommendations", with video platforms (90.5%), shopping platforms (74.4%), social media (67.7%), music platforms (66.1%), and lifestyle consumption software (39.2%) being the platforms frequently used by respondents that contain "algorithm recommendations". Feng Zixuan, Executive Dean and Professor of the Digital Rule of Law Government Research Institute at Southwest University of Political Science and Law, believes that in the digital age of information overload, users find it difficult to obtain beneficial content and information from massive amounts of data. Enterprises use algorithm recommendations to accurately recommend relevant content to users, enhance their user experience, and thus improve the competitiveness of the platform itself. Platforms can also achieve commercial benefits through algorithm recommendations. As a result, various algorithm recommendations are being developed and adopted by more and more platforms. Familiar with the characteristics of algorithm recommendations, Ni Mengna even "domesticates" commonly used apps, labeling herself first to facilitate app recommendations of content she likes. When watching videos on some video platforms, she also specifically uses "one click three links" (a popular internet slang, referring to liking, coins, and bookmarking video works - journalist's note), allowing the app to clearly understand what she likes. If I quit after watching it, the algorithm may not be able to capture my liking for the video, and the content recommended to me may be a bit off track. "Although she does not resist algorithm recommendations, Ni Mengna admitted that she has also encountered" algorithm assassins ". When I ride the same type of car and take the same route with my friends at the same time, my taxi fare will be five or six yuan higher than my friends. The discount coupons given to us by the platform are different. Every time I travel to a city, Liu Haoran will browse local travel guides and store recommendations on short video platforms. However, Liu Haoran did not buy into the carefully prepared content of these algorithms. In his opinion, the content is mostly homogenized, and the copy, camera movements, and music used in video shooting are almost all the same template, with little reference value. Wang Mingming (pseudonym), who is studying at a university in Yunnan, is a radio enthusiast. This niche hobby makes him laugh and cry when it comes to algorithm recommendations. He sometimes purchases electronic components on shopping apps. If the amount of data is large enough, the recommendation will be relatively accurate, but because my hobby is very niche, the algorithm may not be able to capture enough data. For example, when I buy an RF mixer, it recommends a mobile signal amplifier to me, but it has nothing to do with my hobby. "Wang Mingming, who feels uncomfortable with strange and unusual recommendation content, sometimes jokes with the algorithm," I will collect some products that look fun but I will never buy them, confusing the algorithm's visual and auditory aspects. He and his friends also discussed that some software did not have a significant effect after disabling personalized recommendation functions, so they exchanged ideas on how to "reverse algorithms". Wang Mingming's "random collection" method also spread among friends. Algorithm recommendation 'seems to understand me better than me' "because after scrolling through a video of a street side braised food stall on a short video platform, Cheng Huiping, a post-95s from Jiangxi, fell into the" pit "of food and food broadcasting. The platform found that she stayed on such videos for a long time and always recommended her. Now she has followed many food bloggers under the platform's recommendation. Short dramas are also often recommended as a type of video for her. "The final plot of each episode of the short drama is particularly 'gripping', making people unable to resist wanting to watch. At this point, scrolling through two more videos will lead to the following plot. The platform keeps pushing, making people unable to resist scrolling. Dou Yinghui's major is digital media art, and in her daily learning tasks, she needs to use relevant social media platforms to collect materials. She said, "Some image platforms only need to click on one image to push a huge number of related images, greatly improving my efficiency in finding materials." The survey showed that 60.3% of respondents were satisfied with the content recommended by the algorithm. In the eyes of the respondents, the value brought by algorithm recommendations includes easier access to interesting content (68.8%), improved information acquisition efficiency (58.0%), increased opportunities to encounter new things (43.6%), provision of more practical learning resources (26.7%), and matching like-minded friends (24.9%). Two months ago, Yuan Jiale from Zhejiang Ocean University bought a second-hand camera with the help of algorithm recommendation, saving a whopping 1700 yuan. As soon as I placed my order at the original price, I came across a post by a classmate from the same school about second-hand cameras. Many sellers who were closer to me also appeared on the recommendation page, while most of the previous information was from other cities. Yuan Jiale believes that algorithm recommendation also has its advantages: "Algorithm filtering information is more effective, and I am also more likely to find friends who match my interests and hobbies." During the preparation for the International Talent English Exam (Intermediate), Yuan Jiale often collected preparation strategies on social media, and content that met her identity and needs, such as "15 day rapid preparation" and "What steps do non English majors need to take to pass the exam," appeared successively on the recommendation page. The content recommended by the algorithm helped me fill many gaps in my knowledge and methods, and it seems that he understands me better than me. In the end, Yuan Jiale achieved satisfactory results. For Wang Shuo from Chang'an University, watching videos is not only a way to relax, but also an important way to replenish energy and knowledge. He was reviewing professional courses on automotive theory, and after searching for relevant content on video websites, many course videos and learning materials on automotive theory were added to the videos recommended by the platform. Algorithms that are too familiar with individual behavior can also bring problems. When booking a hotel in another city, Cheng Huiping accidentally compared the prices of two different booking apps and found that for the same room type in the same hotel, the app she commonly used was 80 yuan more expensive than the one she rarely used. I have a friend who runs a vacation cabin. He told me that this may be due to 'killing acquaintances', or it may be because the merchant has purchased paid promotion services from the platform. Therefore, the push notifications we see may not only be based on user profiles, but may also be mixed with advertising and marketing. Feng Zixuan admitted that there are many difficulties in the current management and supervision of algorithm recommendations, and there are also problems such as scattered content and unclear standards in relevant regulations; Secondly, algorithm recommendation technology itself has complexity, which makes it difficult for regulatory agencies to have a deep understanding of the working principles and decision-making processes of algorithms, thereby affecting regulatory effectiveness. Algorithm recommendation still needs to be diversified. More than 60% of the surveyed college students have clicked on "not interested" and "algorithms recommend the most concerning and interesting content to users, reducing the areas that users are less involved in, and keeping users in a self circulating information loop. Everyone understands that this" cocoon "has become increasingly strong, but it is difficult to abandon and break," said Dou Yinghui. Recently, Dou Yinghui is in her final week, and "the search keywords are all related to professional courses such as' interactive art design 'and' digital applications'. I also have other hobbies in my life, such as taking photos and watching movies, but due to the lack of relevant information recently, they have slowly faded out of my sight. And fields that I pay less attention to, such as current affairs, politics, mathematics, and science, almost never appear on my social media platforms." Yuan Jiale believes that she relies heavily on algorithms and sometimes finds it difficult to extricate herself from the information flow recommended by algorithms. Time unconsciously passed by my fingertips, and I realized it was already an hour later. "Yuan Jiale would spend more time because he couldn't stop, ultimately" passively staying up late "and harming his body. She has also been trying to find ways to reduce her dependence on algorithm recommended content. The survey shows that the negative effects of algorithmic recommendations in the eyes of respondents include a single source of information (61.7%), homogenization of recommended content (61.7%), and unconscious spending of more time (51.5%). The survey also showed that 72.9% of respondents would actively turn off personalized algorithm recommendations as needed, and 66.6% of respondents would consciously use "not interested" to avoid potential problems caused by "algorithm recommendations". Liu Haoran does not fully accept the information recommended by algorithms, and the comment section of the video is an important source of information for him to assist in judgment. If the information in the comment section is limited, he will search for content posted by other users or switch to other software to obtain more information. The more sources of information, the easier it is to obtain valuable content, "he said. Liu Haoran believes that although algorithm recommendation improves the efficiency of information matching, individuals should have more choice. "When obtaining personal information, the platform needs to confirm with me whether it is allowed, and I can choose whether to accept the content recommended by the algorithm at any time. Ni Mengna hopes that the app can recommend some content to her that she doesn't like, but the proportion should be controlled. "Recommend 80% of the content that I like and 20% of the content that I don't often watch, to avoid being trapped in the 'information cocoon'. Feng Zixuan believes that when ordinary users feel uncomfortable with algorithmic recommendations or believe that data scraping or recommendations have infringed their rights, they can turn off algorithmic recommendation services on their own. Users should enhance their self-protection awareness when using platform services. For example, they should pay attention to protecting personal privacy and information security, and carefully authorize the platform to obtain their personal information and data. Feng Zixuan said that if users are dissatisfied with algorithm recommendations or believe that their rights have been infringed, they can file a complaint with the platform and seek solutions. When the rights and interests of users are seriously infringed and the platform fails to provide a reasonable solution, users can consult lawyers or relevant institutions to protect their legitimate rights and interests in accordance with the law. What does a good algorithm recommendation look like? In the eyes of the respondents, considering user diversity needs (79.1%), increasing filtering of pushed content (59.3%), allowing users more control (58.3%), ensuring user privacy and data security (46.4%), explaining the recommendation mechanism clearly to users (40.1%), and prioritizing the recommendation of positive and beneficial content (37.1%) are all indispensable factors. Sometimes the problems brought by algorithm recommendations are not something that ordinary users can decide or influence, and must be solved from the root by the platform and relevant regulatory departments Cheng Huiping said. However, she will still consciously discern platform algorithm recommendations based on her experience