As the era of large-scale models approaches, the data annotation industry is experiencing a new wave of enthusiasm. However, the plight of data annotators has not improved; instead, they face more severe challenges. Through interviews with multiple data annotators, the article reveals that their primary concerns are low wages, high turnover, and a sense of meaningless and monotonous work, often feeling underestimated by algorithm engineers. Meanwhile, AI companies are increasingly adopting automated annotation and synthetic data technologies to replace manual annotation, putting annotators at risk of being phased out. Data shows that 70% of the foundational data used for artificial intelligence abroad is now synthetic. The article calls for attention to the difficulties faced by data annotators and emphasizes the need to consider the rights of grassroots workers while advancing AI development.