(Source: CCTV13 2025-02-20)

Recently, a set of data on the mortality rate of post-80s has attracted widespread attention online.CCTV reporters found that there are many articles on the Internet about the mortality rate of the post-80s generation that quoted this data, and many self-media articles are constantly exaggerating, such as: "The 5.2% mortality rate of the post-80s generation has opened up the sad departure of the golden generation"; for example: "When the post-80s generation begins to leave collectively - Behind the 5.2% mortality rate, the survival dilemma of a generation and the sorrow of the times." There are also online pictures that make the so-called mortality rate of post-80s a bar chart, saying: "One of every 20 people has passed away", and labeled the comparison data, saying that "the mortality rate of post-70s is lower than that of younger post-80s."

The source of the erroneous mortality rate may be "AI operation deviation"

Where does this sensational and loophole-filled mortality data come from? Li Ting, a professor at the School of Population and Health of Renmin w88 online casino betting of China, found that it is very likely that the AI ​​model has an error in the answering of the question.

For the hallucination problems and generation mechanism of the big model,Professor Zhang Xin of Law SchoolAfter a telephone interview with CCTV13 News Channel, we analyzed the operation mechanism of the big model, the reasons for the large model illusion in the data layer, model layer and application layer, the current network information governance ideas in my country, and the legal liability that may be caused when deliberately spreading false information causes adverse effects.

For the technical causes of "illusion" of the big model,Zhang XinNot that the data layer, model layer and application layer may cause "illusion" of big models. At the data level, the timeliness deviation and quality defects of the training corpus are the primary incentives, especially after the false information implicit in the data crawled on the network is processed by the model parameterization, it is easy to form systematic cognitive bias.

my country has previously issued a series of departmental regulations represented by the "Regulations on Deep Synthesis Management of Internet Information Services" and the "Interim Measures for the Management of Generative Artificial Intelligence Services". Recently, it has also been speeding up the construction and improvement of artificial intelligence generated synthetic content identification methods and standards to provide a clear compliance framework for generation synthetic service providers and content dissemination service providers. If false and harmful information is deliberately fabricated, it may also bear corresponding criminal liability if it is spread through the Internet to cause public panic, disrupt public order, and cause adverse social impacts.

Reporter found that after the Internet fermentation, the mortality rate has become a gimmick for some self-media to sell goods.It can be seen that although AI models have super strong data storage and computing capabilities, they may lead to incorrect calculations due to insufficient training corpus and limited data sources. In order to monetize traffic, some self-media will take things out of context and use them to promote the hype and dissemination of these outrageous data.

Attachment: Original link

https://content-static.cctvnews.cctv.com/snow-book/index.html?item_id=2084160556267115271&t=1740056794180&toc_style_id=feeds_default&share_to=wechat&track_id=23ae2058-645c-47f2-b968-4321ae119902