Surge AI

Edwin Chen founded Surge AI in 2020, inspired by dissatisfaction with the quality of crowdsourced data labeling. Early projects had focused on search, recommendation engines, and content moderation. For data labeling, they developed a platform to match annotators to their expertise. These annotators reviewed the responses by generative AI models, including chatbots and text-to-image models, improving the model outputs by sharing information reflecting the annotators' expertise. In 2025, the company worked with about 1 million annotators. {{Cite web |last=Rollet |first=Charles |date=2025-09-25 |title=AI training companies are raising billions to get humans to teach chatbots. Here are the startups cashing in. == Controversies ==

The Verge and New York Magazine reported in 2023 that Surge AI appeared to own multiple separate platforms for work, including Get Hybrid, Task Up, and Data Annotation. and unexplained annotator account cancellations. In May 2025, Surge AI faced a class action lawsuit that the company "deliberately" misclassified its data annotators as independent contractors, denying them employee benefits and improperly withholding wages. In July 2025, two documents from Surge AI had been leaked, one on model training and safety guidelines, and another on websites that contractors training Anthropic models were and were not allowed to use. == References ==