【环球时报驻美国特约记者 卓然 环球时报记者 李迅典 环球时报特约记者 刘皓然】 编者的话： 人工智能（AI）的高速发展正重塑东西方的就业格 局：在取代部分低门槛、重复性劳动的同时，也在创造新的就业机会。就在美国科技巨头忙于降本裁员之际，AI相关产业的外溢却盘活了印度 的"三线小城"，不少"小镇青年"走进写字楼，成了 "白领职员"。而在美国，一些焦虑的"00后"为了不被AI取代，转向电工、水管工等"技术工 种"。这种反差现象的背后，其实是资本对于用工成本的精密考量——当美国"大厂精英"供不起时，印度"小镇青年"成为更具性价比的选择。 AI 热潮带动印度小城就业 在印度AI数据公司Objectways的一间办公室，数据标注员达拉妮紧盯着屏幕上的一组人物对比图。很快，她根据面目特征的异常分辨出假图、完 成标记，并继续进行下一组识别。在办公室的另一端，她的同事索米娅进行着视频识别工作，在每个视频片段播放完成后，她需要标记出这段视 频内容的类别、视频中人物的语气与情绪，帮助AI大模型进行"学习"。 《日经亚洲评论》报道称，AI数据标注员并非新生事物，但随着技术的不断发展，行业对标注员的工作要求也在更新。报道援引Obj ...

Core Insights - Surge AI, founded in 2020, has achieved $1 billion in revenue without any external funding, positioning itself as a significant player in the AI data annotation industry, surpassing competitors like Scale AI, which generated $870 million in revenue and has raised $1.6 billion in funding [2][3][4]. Group 1: Company Overview - Surge AI has a team of around 120 people and counts major companies like Google, OpenAI, and Anthropic as clients [2]. - The company focuses on delivering high-quality data specifically for training and evaluating AI models, contrasting with competitors that primarily offer human outsourcing services [8][18]. Group 2: Business Philosophy - The founder, Edwin Chen, emphasizes that entrepreneurship should focus on solving problems rather than seeking funding, highlighting that the current hype around synthetic data is overestimated [5][9][12]. - Surge AI's business model is built on the belief that high-quality human data is essential for AI development, as opposed to relying on synthetic data, which often proves ineffective in real-world applications [11][44]. Group 3: Data Quality and Challenges - Surge AI differentiates itself by prioritizing data quality, employing complex algorithms to ensure the data provided is of the highest standard, unlike many competitors who lack technological capabilities [20][26][34]. - The company recognizes the challenges in maintaining data quality, noting that even highly educated individuals may produce subpar data if not properly managed [21][24]. Group 4: Market Trends and Future Outlook - The discussion around synthetic data reveals that it is often inadequate for training models effectively, with many clients realizing the limitations after extensive use [45][49]. - The future demand for diverse data types, including reinforcement learning environments, is expected to grow, as models require more complex and varied inputs to perform well [37][43]. Group 5: Evaluation Standards - Human evaluation is deemed the gold standard for assessing model performance, as it allows for a more nuanced understanding of quality beyond superficial metrics [76]. - Surge AI aims to promote a deeper understanding of model capabilities and limitations, advocating for thorough human assessments rather than relying on quick, subjective evaluations [77].