AI News AI资讯 4d ago Updated 4d ago 更新于 4天前 42

Amazon sunsets Mechanical Turk, the original "Artificial Artificial Intelligence" 亚马逊关停机械土耳其,最初的“人工人工智能”

AWS will cease accepting new customers for Mechanical Turk, SageMaker Ground Truth, and Amazon Augmented AI starting July 30, 2026. Existing users may continue using these services, but they are entering maintenance mode with no new feature development. The shutdown reflects the declining utility of general crowdsourcing for high-quality AI training data due to the rise of LLMs. Industry preference has shifted toward specialized vendors like Scale AI and Surge AI for expert-level data annotation AWS宣布将于2026年7月30日起停止向新客户开放Mechanical Turk、SageMaker Ground Truth及Amazon Augmented AI服务。 现有用户可继续使用,但服务将进入维护模式,不再推出新功能。 众包平台价值因大量工人使用语言模型生成数据而受到削弱,导致其作为人类数据源的地位下降。 AI实验室正转向Scale AI或Surge AI等拥有经过验证专家的专业供应商,以处理复杂边缘案例和质量判断。

65
Hot 热度
60
Quality 质量
55
Impact 影响力

Analysis 深度分析

TL;DR

  • AWS will cease accepting new customers for Mechanical Turk, SageMaker Ground Truth, and Amazon Augmented AI starting July 30, 2026.
  • Existing users may continue using these services, but they are entering maintenance mode with no new feature development.
  • The shutdown reflects the declining utility of general crowdsourcing for high-quality AI training data due to the rise of LLMs.
  • Industry preference has shifted toward specialized vendors like Scale AI and Surge AI for expert-level data annotation.

Why It Matters

This marks the end of an era for general-purpose human-in-the-loop data collection, signaling that broad crowdsourcing platforms are no longer sufficient for modern AI development needs. For practitioners, it highlights the critical importance of data quality and the necessity of partnering with specialized providers who offer vetted expertise rather than generic labor.

Technical Details

  • Service Sunset Timeline: New customer onboarding for Mechanical Turk, SageMaker Ground Truth, and Amazon Augmented AI stops on July 30, 2026.
  • Maintenance Mode: Legacy users retain access but receive no further updates or feature enhancements.
  • Historical Context: Mechanical Turk launched in 2005 as "Artificial Artificial Intelligence," evolving from general micro-tasks to AI data annotation tools by 2018.
  • Quality Degradation Factors: A 2023 study indicated that crowdworkers increasingly used language models to perform tasks, compromising the integrity of human-generated ground truth data.

Industry Insight

  • Consolidation of Data Vendors: The market is consolidating around specialized players (e.g., Scale AI, Surge AI) that provide higher assurance of data quality and expert handling of edge cases.
  • Strategic Shift in Data Acquisition: Organizations should prioritize building relationships with niche data providers over relying on mass-market crowdsourcing platforms for training robust models.
  • Legacy System Planning: Teams currently using AWS data annotation tools must begin migration planning well before 2026 to avoid service disruption.

TL;DR

  • AWS宣布将于2026年7月30日起停止向新客户开放Mechanical Turk、SageMaker Ground Truth及Amazon Augmented AI服务。
  • 现有用户可继续使用,但服务将进入维护模式,不再推出新功能。
  • 众包平台价值因大量工人使用语言模型生成数据而受到削弱,导致其作为人类数据源的地位下降。
  • AI实验室正转向Scale AI或Surge AI等拥有经过验证专家的专业供应商,以处理复杂边缘案例和质量判断。

为什么值得看

本文揭示了人工智能数据标注基础设施的重大变迁,标志着通用众包模式在高质量AI训练数据需求面前的退场。对于AI从业者和企业而言,这明确了从廉价劳动力依赖向专业化、专家级数据服务转型的行业必然趋势。

技术解析

  • 服务终止时间表:AWS通过服务可用性更新公告,确定2026年7月30日为Mechanical Turk及相关AI数据工具停止接纳新客户的截止日期。
  • 历史演变与失效原因:Mechanical Turk自2005年推出以来曾定位为“人工人工智能”,但在2018年试图转型为AI训练数据标注工具后,2023年的研究指出众包工人广泛使用LLM生成数据,严重破坏了其提供真实人类标注数据的价值。
  • 替代方案对比:当前行业倾向于使用Scale AI或Surge AI等专业供应商,这些公司雇佣经过严格筛选的专家,能够更有效地解决数据标注中的边缘情况(edge cases)和质量控制问题。

行业启示

  • 数据质量门槛提升:随着基础大模型能力的增强,对训练数据中“人类智慧”和“专家判断”的需求日益稀缺且昂贵,通用众包无法满足高质量对齐数据的需求。
  • 供应链专业化重组:AI数据服务市场正在经历洗牌,低附加值的批量众包将被高附加值的专业化、专家型数据服务所取代,相关企业需调整数据采购策略。
  • 长期规划必要性:依赖AWS等云平台数据服务的团队需提前规划数据标注迁移路径,特别是在2026年截止日期前完成向专业供应商的过渡。

Disclaimer: The above content is generated by AI and is for reference only. 免责声明:以上内容由 AI 生成,仅供参考。

Closed Source 闭源 Product Launch 产品发布 Policy 政策