刚刚过去的2025年，被业界称为算法透明"元年"。这一年，短视频、外卖、网约车及货运平台纷纷公开 算法模型和原理、平台调度规则，回应网友和社会关切。随着互联网平台的算法治理走向深入，算 法"黑箱"正在向"白箱"转变。 来源：科技日报 ◎ 杨雪 需要正视的是，在全球范围内，算法治理和技术公开透明化是一个普遍难题，面临技术复杂性、商业利 益平衡及法律框架完善等多重挑战。构建健康、普惠的网络生态，未来还需多方协同发力。除了互联网 平台主动作为，监管机构应实施常态化监管，社会公众也要积极参与治理过程，把"人"的诉求尽可能多 地反映给"机器"。 人工智能时代，算法作为一种先进工具，日益嵌入日常生活的方方面面。它显著提升线上交互效率，改 善用户体验，助力平台经济繁荣发展。但与此同时，"大数据杀熟""骑手内卷"等算法应用带来的弊端， 也日益受到诟病。算法如何推荐商户？如何分配订单？如何预估配送时间？——从舆论热议的"外卖大 战"可以清楚看出，算法的设计稍有偏差，各方的利益关系就可能失去平衡，带来社会矛盾。 算法治理不易，"元年"只是起点，而我们已经看到算法治理推动技术向善的清晰方向。 算法透明是公众对人工智能的信任基础。让 ...

◆全国总工会与最高检等部门在保障劳动者权益方面进行了诸多制度创新和实践创新，包括构建"工会 +法院+检察院+人社+司法"劳动争议多元化解机制，运用"一函两书"等。 ◆广大劳动者切身利益具有公共利益属性，应将劳动者权益保障写入检察公益诉讼法。 在数字技术重塑劳动关系的当下，"算法"已成为平台企业管理的核心工具，其背后的劳动者权益保障问 题日益凸显，亟待解决。 全国总工会开展的第九次职工队伍状况调查显示，全国新就业形态人员约8400万人。如此庞大的群体， 如何为其构建完善的权益保障体系，已成为立法、司法、行政部门与工会共同面临的重大课题。 2025年岁末，在上海市青浦区检察院与区总工会的协同推进下，部分快递物流企业率先完成算法治理。 2025年12月25日，全国总工会、最高法、最高检联合发布"2025年劳动法律监督'一函两书'典型案例"， 该案入选，成为破解算法困局的实践样本。 2025年11月26日，北京市东城区崇文门外街道"崇外e家"的工作人员为外卖配送员提供应急药品。新华 社记者李欣摄 ◆现行劳动关系矛盾调处机制不能完全适应劳动争议矛盾量大、难以覆盖新就业形态等要求。 困局：新业态给劳动争议解决带来新难题 ...

Core Viewpoint - The release of the "Antitrust Compliance Guidelines for Internet Platforms (Draft for Comments)" by the State Administration for Market Regulation represents a significant step towards enhancing the antitrust regulatory framework for platform economies, aiming to promote healthy development through compliance [1] Group 1: Ecological Characteristics of Platform Economy - The guidelines emphasize the ecological nature of platform economies, highlighting the need to establish a healthy competitive ecosystem [2] - It defines platform operators as ecosystem managers, responsible for maintaining positive relationships within the platform ecosystem, thus making ecosystem maintenance a key antitrust regulatory goal [2] - The complexity of antitrust governance in platform economies is underscored, as it must consider the rights of various stakeholders within the ecosystem, not just the competitive actions against market players [2][3] Group 2: Technological Aspects of Platform Economy - The guidelines recognize the technological characteristics of platform economies, particularly the role of advanced technologies like cloud computing and AI in driving growth [4] - It identifies the dual nature of technology as both a facilitator of market efficiency and a potential tool for anti-competitive behavior, thus establishing clear compliance boundaries for platform operators [5] - The guidelines encourage platforms to enhance algorithm governance and implement comprehensive compliance management to mitigate algorithm misuse risks [6] Group 3: Proactive Measures Against New Antitrust Risks - The guidelines address the challenges of defining and preventing new forms of monopolistic behavior, providing actionable guidance on issues like "algorithm collusion" and "blocking" [7] - It highlights the risks of irrational competition in platform economies, particularly through excessive price competition and the imposition of unreasonable conditions on merchants [8] - The guidelines clarify the compliance boundaries regarding practices like "blocking" and "differential treatment," aiming to foster a more unified digital market and stimulate innovation within the platform economy [9]

Core Points - The core focus of the meeting was on "algorithm for good," addressing concerns of truck drivers regarding platform algorithms and labor rules [1][2][3] Group 1: Meeting Overview - The meeting was a follow-up to the first national democratic consultation held by Zhongchu Zhiyun, aimed at engaging with new economy truck drivers [1] - Key attendees included representatives from various labor unions and government departments, indicating a collaborative effort to address driver concerns [1][3] Group 2: Algorithm and Driver Concerns - Truck drivers raised questions about the algorithm used for dispatching orders, specifically whether it favors drivers with longer online times [2] - The platform clarified that the dispatch algorithm prioritizes matching efficiency rather than online duration, aiming for fairness in order allocation [2][4] Group 3: Outcomes and Agreements - The meeting resulted in the signing of the "Zhongchu Zhiyun Platform Algorithm and Labor Rules Agreement," which formalizes the drivers' collective demands into a regulatory document [2][5] - The agreement includes commitments to address issues such as labor compensation, safety, and service guarantees, reflecting a deeper engagement compared to previous discussions [2][3] Group 4: Future Collaboration and Mechanisms - Zhongchu Zhiyun plans to hold at least one formal democratic consultation annually, establishing a continuous dialogue with truck drivers [5] - The initiative is seen as a model for algorithm governance and labor rights protection, promoting a collaborative relationship between the platform and drivers [5]

Core Viewpoint - The article discusses significant changes in the jurisdiction of internet courts in China, effective from November 1, focusing on complex cases while removing simpler, repetitive cases from their purview [3][5]. Summary by Sections Changes in Jurisdiction - The Supreme Court has issued new regulations that expand the jurisdiction of internet courts in Beijing, Hangzhou, and Guangzhou to include four new types of cases: "network data," "personal information protection and privacy," "virtual property," and "unfair competition disputes" [3][5]. - Simultaneously, it removes simpler cases such as "financial loans," "small loan contract disputes," and certain "copyright disputes" from its jurisdiction, addressing the issue of overwhelming simple cases that consume judicial resources [3][5][6]. Historical Context and Issues - Initially, internet courts were established to explore online litigation rules, but the effectiveness of this mission has diminished as many courts can now conduct online hearings [6][10]. - The Beijing Internet Court reported handling 194,000 cases from its establishment in 2018 to July 2023, with a high concentration of simple cases, leading to a "case overload" issue [6][10]. Strategic Repositioning - Experts suggest that the recent adjustments are not merely minor tweaks but a strategic reconfiguration of the internet courts' functions, aiming to focus on complex and emerging digital economy-related disputes [6][9]. - The article emphasizes the need for dynamic adjustments in jurisdiction based on the maturity of case types and the necessity for rule exploration in new areas [7][9]. Future Directions - The internet courts are expected to evolve into "rule incubators" for governance in the digital space, addressing issues like data rights, platform economy, and algorithm governance [14][15]. - There is a call for establishing judicial review standards for platform rules, ensuring fairness and transparency in algorithmic decision-making [15][16]. - The article highlights the importance of maintaining judicial integrity while innovating procedural rules, ensuring that the rights of parties involved are not diminished [16].

Core Viewpoint - The Supreme People's Court has issued new regulations to enhance the jurisdiction of internet courts, effective from November 1, 2025, to better serve the digital economy and improve online dispute resolution [1][2]. Group 1: Background and Reasons for Regulation Adjustment - The adjustment of jurisdiction is a response to the need for improved governance in the digital economy and aligns with the directives from the 20th Central Committee's Third Plenary Session [3]. - The changes are necessary to fulfill the functional positioning of internet courts as outlined by the Central Committee, focusing on online dispute resolution and legal governance in the digital space [4]. - The adjustment addresses the diverse judicial needs of the public in the digital age, as new types of legal disputes related to data, algorithms, and personal information protection are emerging [5][6]. Group 2: Impact on Litigation - The new regulations will change the jurisdiction of specific cases, such as data rights and online personal information disputes, which will now be handled by internet courts [7]. - The adjustment aims to enhance the litigation experience for parties involved, ensuring that they can still benefit from online dispute resolution mechanisms [8]. Group 3: Judicial Capacity and Resources - Internet courts have accumulated substantial experience and resources to handle new types of network-related cases, having previously dealt with significant cases in this domain [9]. - The courts have established a specialized team of judges with expertise in both law and technology, ensuring effective handling of complex digital disputes [9]. Group 4: Broader Implications for Internet Courts - The jurisdictional changes will enhance the overall level of internet adjudication across the country, serving as a model for other courts [10]. - The adjustments will help establish clearer judicial standards for emerging fields, thereby strengthening the legal framework for digital economy governance [11]. - The regulations will also expand the jurisdiction to include international network cases, promoting better engagement in global internet governance [12].

Core Points - Meituan announced the complete cancellation of "overdue fines" for delivery riders by the end of 2025 [1][3] - The initiative is part of Meituan's broader algorithm governance measures aimed at improving rider experience and income stability [4][5] Group 1: Cancellation of Overdue Fines - The cancellation of overdue fines will be implemented nationwide after successful trials in 22 cities, including Hangzhou and Wuhan [5] - Meituan has been exploring positive incentives to replace fines, with a scoring system to reward high-performing riders [4][6] - Feedback from riders indicates a significant improvement in their experience, as they no longer fear penalties for delays [4] Group 2: Enhancements in Delivery Efficiency - Meituan is collaborating to establish "rider-friendly communities" to address common delivery challenges, with over 24,700 communities modified to facilitate rider access [6] - The company is implementing measures to improve address accuracy and user location sharing to reduce delivery issues [6] - A "safety score" system has been introduced to incentivize safe riding behavior, resulting in a 26% reduction in traffic violations among participating riders [6] Group 3: Commitment to Transparency and Fairness - Meituan's public affairs representative emphasized the importance of transparency and collaboration in optimizing delivery algorithms [7] - The company aims to continuously refine its management practices to ensure fairness and humanity in rider treatment [5][7]

Group 1 - Meituan announced the complete cancellation of the "overtime penalty" for delivery riders by the end of 2025, addressing a major concern among the crowd-sourced riders [1] - The company has implemented a new system called "Anzhun Card" to replace the previous penalty system, which has been piloted in 22 cities and received positive feedback from riders and the public [1][2] - Meituan is shifting its management approach from punitive measures to positive incentives, focusing on training and learning instead of solely penalizing riders [1] Group 2 - The company has introduced fatigue prevention measures, including pop-up reminders after 8 hours of work and mandatory offline after 12 hours, with 18% of riders triggering the 8-hour reminder [2] - Meituan has launched a "safety score" system in over 100 cities to improve traffic safety among riders, resulting in a 26% reduction in traffic violations [2] - The company plans to enhance the reporting and appeal processes for riders using AI to improve efficiency and quality in problem resolution [2] Group 3 - In Q2, Meituan reported revenue of approximately 91.84 billion yuan, an increase of 11.7% year-on-year, while adjusted net profit decreased by 89% to approximately 1.493 billion yuan [3] - The core local commerce segment saw a revenue increase of 7.7% to 65.3 billion yuan, but operating profit dropped by 75.6% to 3.7 billion yuan due to irrational competition [3] - New business revenue grew by 22.8% to 26.5 billion yuan, with operating losses expanding by 43.1% to 1.9 billion yuan, although the operating loss rate improved by 3.1 percentage points [3]

Core Insights - Meituan held an open day for rider algorithm co-governance, revealing the logic behind its delivery algorithms and sharing improvements in rider work experience [1] - The company plans to eliminate overtime penalties for riders by the end of 2025, addressing a major concern among crowd-sourced riders [1] - A pilot program for the no-penalty mechanism has been implemented in 22 cities, receiving positive feedback from riders and the public [1] Group 1 - The open day is the eighth communication event since the announcement of eight algorithm governance measures and the establishment of an algorithm advisory committee earlier this year [1] - The company has conducted trials in over ten cities over the past eight months to compare different management models, ensuring stable rider income and user experience [1] - Positive incentives are provided for high-performing riders, enhancing their overall experience [1] Group 2 - The no-penalty assessment mechanism will be piloted in first-tier cities, including Hangzhou and Wuhan, starting July 2025, with plans for gradual expansion [1] - Riders in regions like Guangxi and Jiangxi have reported improved experiences, both for new and long-term part-time crowd-sourced riders [1] - The initiative aims to address challenges such as the "last mile" delivery difficulties and enhance safety and fatigue management for riders [1]

Core Viewpoint - The article discusses the phenomenon of "big data price discrimination" where consumers face higher prices based on their purchasing behavior and history, leading to a growing backlash and various strategies from users to combat this issue [1][2][3]. Group 1: Consumer Experiences - Consumers have reported experiencing price discrepancies for the same products based on their shopping frequency, with examples of individuals finding price differences of 10 yuan for the same item when using different accounts [2]. - Common forms of "big data price discrimination" include automatic price increases after multiple views and different prices for the same product at the same time for different users [2][3]. Group 2: Legal and Ethical Considerations - The distinction between "big data price discrimination" and legitimate "differential marketing" is legally ambiguous, requiring clearer guidelines for recognition [3][9]. - Legal frameworks have been established to protect consumers' rights against "big data price discrimination," emphasizing the need for algorithm governance to ensure fair practices [9][10]. Group 3: User Strategies - Users are employing various tactics to counteract "big data price discrimination," such as creating multiple accounts, using different devices, and posting reverse comments to signal financial constraints [4][6]. - The effectiveness of these strategies is debated, with some users reporting temporary success, while others find that such tactics do not consistently yield lower prices [6][7]. Group 4: Industry Response and Governance - Recent initiatives by platforms aim to address "big data price discrimination," with companies like Pinduoduo and Tencent announcing measures to enhance algorithm transparency and consumer protection [9][10]. - The article highlights the need for a collaborative approach among government, businesses, and consumers to establish a fair digital economy and mitigate the impacts of "big data price discrimination" [10].