Core Viewpoint - The incident involving DeepSeek's alleged apology for associating Wang Yibo with the "Li Aiqing corruption case" highlights significant issues in AI model reliability and the propagation of false information in the digital age [3][6]. Group 1: Incident Overview - DeepSeek reportedly apologized for incorrectly linking Wang Yibo to a corruption case, citing a content review oversight and the use of unverified rumors [3]. - Despite widespread media coverage claiming an apology, no official statement or acknowledgment from DeepSeek was found on any of its platforms [3][4]. - The incident illustrates a broader issue where AI models, when queried about the event, consistently confirmed the existence of an apology based on the same false news sources [4]. Group 2: AI Model Challenges - The situation reveals that AI models face a more severe challenge than the commonly discussed "hallucination problem," as they can generate new false news by referencing existing false information [6]. - Although advancements in AI have reduced the hallucination problem, the prevalence of false information online complicates the models' ability to verify facts effectively [6]. - The "Rubbish in, Rubbish out" effect is evident, where the quality of AI-generated content is directly influenced by the accuracy of the input data [6]. Group 3: Implications for Human and AI Interaction - The experience of news professionals indicates that AI cannot fully replace human writers, as verifying AI-generated information often consumes more time than traditional writing methods [7]. - The incident serves as a warning for both AI and humans regarding the importance of discerning truth in a landscape filled with misinformation [8]. - Companies utilizing AI models are encouraged to integrate proprietary knowledge bases to enhance the accuracy of AI-generated responses [7].