新京报讯（记者秦胜南）4月25日，2025年度"饿了么"（上海）协商算法恳谈会上，形成《2025年度"饿 了么"（上海）网约配送算法和劳动规则协议》，在全国总工会、上海市总工会、普陀区总工会的共同 见证下，由外卖送餐员代表与平台企业代表签字确认生效。据悉，这也是行业首份网约配送算法和劳动 规则协议。 这一协议关乎骑手切身利益的方方面面，聚焦计价、调度派单、时间预估、路径规划等"四大算法"模型 及规则向上向善、公平透明，签订协议涵盖总则、劳动报酬、休息休假、劳动保护、协商协调机制五个 方面、三十三条。 恳谈会上，围绕计价算法，9名骑手代表带来一线小哥最为关注的议题。经过各方协商达成：平台计价 算法方面，优化"恶劣天气自动触发补贴"功能；平台建立平台算法和劳动规则民主公示制度，接单页面 中，明确标识配送费计算规则（如距离、时段、难度系数等）。 针对强制休息规则，经过多轮协商，多方达成一致，平台全面升级防疲劳规则，新增"自主选择休息权 提醒"功能、骑手可自行选择期望的休息时间段，增加灵活调整空间，自助选择休息时间，降低配送疲 劳度。 此外，在配送时长计算中，各方达成一致，平台将优化时间预估算法模型及规则，优化安全限 ...

Core Viewpoint - Meituan is implementing a traffic safety governance pilot program for delivery riders in several cities, incentivizing safe driving behavior through financial rewards and a safety scoring system [1][2][6] Group 1: Pilot Program Details - The pilot program includes cities like Jinan, Quanzhou, Huizhou, and Baotou, where riders can earn the "Wait for the Light Award" for maintaining a zero red-light running record, potentially increasing their weekly income by over 400 yuan [1] - Over 20,000 riders participated in the pilot, resulting in a 26% reduction in red-light violations among delivery riders, while on-time delivery rates remained stable [1][2] - The program will be expanded to over 100 cities, incorporating a safety score into rider assessments, with higher scores leading to greater rewards and benefits [1][2] Group 2: Safety Initiatives and Community Engagement - Meituan is focusing on a comprehensive traffic safety governance system that includes preemptive training, real-time risk alerts, and post-violation management [6] - The company is collaborating with local traffic police to enhance safety measures, including the "Lighthouse Plan," which connects data with 39 city traffic departments to identify high-risk riders [6] - Riders who maintain safe driving records can accumulate safety points, leading to monthly rewards and additional benefits, while violations can result in penalties such as warnings or restrictions on order acceptance [2][6]

Core Viewpoint - The rise of AI agents, such as OpenAI's Operator, poses potential risks to consumer-facing applications like DoorDash, as these agents could automate tasks traditionally performed by users, potentially bypassing the platforms altogether [3][4][10]. Group 1: Impact on Consumer Applications - OpenAI's Operator, in collaboration with DoorDash, aims to automate consumer tasks, which could lead to a decline in direct user engagement with DoorDash's website [1][3]. - If AI agents become proficient, they may act as intermediaries, reducing the value of advertising on platforms like DoorDash, which relies on restaurant ads as a growing revenue source [3][4]. - Concerns have been raised by DoorDash executives about the potential negative impact of AI agents on their business model, particularly regarding advertising revenue [2][5]. Group 2: Retailers' Response Strategies - Retailers, including Walmart, are considering building their own agents to interact with consumers, thereby maintaining control over product recommendations and information [7][12]. - The potential for AI agents to disrupt traditional advertising strategies has led to discussions among retailers about adapting their approaches to consumer engagement [12][10]. - OpenAI has indicated that AI agents could serve as valuable traffic sources for retailers, suggesting a collaborative rather than purely competitive relationship [5][6]. Group 3: Current Trends and Data - A recent Adobe report highlighted that 39% of surveyed consumers have utilized generative AI for online shopping tasks, indicating a growing trend in AI-assisted consumer behavior [9]. - Despite the rapid growth of AI applications, the current traffic generated by these tools to retail websites remains moderate compared to traditional methods [9]. - OpenAI's ChatGPT is becoming an increasingly important source of referral traffic for retailers, complicating their relationship with AI technologies [10][8].