Workflow
AI检测
icon
Search documents
Wind风控日报 | 比亚迪李云飞回应“汽车圈恒大”质疑
Wind万得· 2025-05-30 22:55
Macro Insights - The 2025 Lujiazui Forum will announce several major financial policies, with key speeches from the Governor of the People's Bank of China and the Deputy Governor [3] Bond Market Alerts - Jilin Credit Financing Guarantee Investment Group disclosed three major litigation cases involving financial loan contract disputes, indicating that the company is operating normally and its financial status is good [5] - Zhaoyang Urban Construction Investment Co., Ltd. is listed as an executor in two cases due to loan contract disputes, with liabilities totaling 65 million yuan [6][14] - China Huayang Economic and Trade Group failed to pay due on bonds, leading to substantial default, with the bankruptcy administrator announcing criminal charges against the issuer [8] - Gansu Energy Chemical's coal mine experienced a flooding accident, resulting in three fatalities, raising concerns about safety management and profitability due to falling coal prices [9] - Xi'an New District's project revenue significantly underperformed expectations, with actual income of 8.1 million yuan against a forecast of 260.2 million yuan, although it has not adversely affected the company's debt repayment ability [10] Stock Market Alerts - The China Securities Index Company announced regular adjustments to the CSI 300 and other indices, with several companies being added to the indices [16] - BYD's management refuted claims that it is akin to "Evergrande in the auto industry," asserting that mainstream Chinese automakers have better asset-liability situations compared to foreign counterparts [19] - ST Dongshi is under investigation by the China Securities Regulatory Commission for suspected information disclosure violations [20] - ST Tianyu's actual controller is under investigation for fund misappropriation, but it is not expected to significantly impact the company's operations [21] - ST Jinglan is also under investigation for information disclosure violations, with no major impact on daily operations anticipated [22] - ST Yazhen warned of potential stock price volatility after a significant increase of 93.14% in its stock price [23] Industry Alerts - Major pig farming companies have received notifications to suspend the expansion of breeding sows and control the weight of pigs for sale, indicating regulatory tightening in the industry [41] - The Ministry of Industry and Information Technology reported a 10.8% year-on-year decline in profits for internet enterprises from January to April, despite a 1.5% increase in internet business revenue [42]
《滕王阁序》AI生成率竟达100%,高校AI检测逼疯师生
3 6 Ke· 2025-05-19 23:45
(一)"数据库依赖症"引发误判连锁反应 当前国内外AI检测系统主要依赖公开资料库及联网搜索比对机制,而经典文学作品因其广泛传播和高度引用,极易被判定为"与现有内容高度重合"。例 如,朱自清的《荷塘月色》因长期收录于教材、文献,其片段被系统标记为"高度疑似AI生成"(AI率62.88%)。 当"AI率"检测系统从"学术辅助者"异化为"新型学术税"的征收者,当"反AI文本"成为一门暴利生意,学术诚信的底线正被技术漏洞与人性贪婪 共同撕裂。在这场荒诞的猫鼠游戏中,学生、商家、平台三方围绕"AI率"展开一场"技术对抗技术"的黑色狂欢,而学术尊严与教育的本质,早 已沦为这场闹剧的"牺牲品"。 当前,毕业论文季的硝烟弥漫全国高校。 复旦大学、西南大学、中国传媒大学等十余所高校相继出台新规,将"AI生成内容检测率"(AI率)纳入论文评审硬性指标,要求本科论文AI率不得超过 15%、硕士论文不超过10%、博士论文需低于5%。这场以"技术监管"为名的学术变革,却在执行中演变为一场"人机对决": 学生被迫与AI"斗智斗勇",用方言插入、错别字掩护、甚至手写扫描再转录等"土办法"降低AI率;教师沦为"算法的执行者",在"降AI率" ...
朱自清《荷塘月色》也是AI代写?网友质疑AI检测科学性 记者实测
Yang Zi Wan Bao Wang· 2025-05-10 07:41
Group 1 - The core issue revolves around the reliability of AI detection systems, as classic literary works like Zhu Ziqing's "Lotus Pond Moonlight" and Liu Cixin's "The Three-Body Problem" are flagged with high AI generation probabilities, raising doubts among netizens about the accuracy of these tools [1][2] - A practical test conducted by reporters showed that the AI generation probability for "Back Shadow" was 18.21%, while "Ball Lightning" had a probability of 32.05%. Other AI detection websites reported even lower rates for "Back Shadow," with some showing less than 1% [2][4] - AI detection models are based on training with large datasets of both human and AI-generated texts, but they can misjudge due to the evolving nature of AI-generated content. Continuous updates to detection models are necessary to reduce misjudgment rates [4][6] Group 2 - The discussion on social media regarding the high AI detection rates in academic papers suggests that educators should not rely solely on these results as a measure of student performance, but rather consider them as part of a broader evaluation system [6]