News Summary

DeepSeek, a prominent Chinese AI lab, has released its latest AI model, DeepSeek V3, which is reportedly outperforming many competitors in popular benchmarks. However, the model has stirred controversy as it identifies itself as ChatGPT, OpenAI’s chatbot platform, and provides instructions related to OpenAI’s API. This has led to speculation that DeepSeek V3 may have been trained on outputs from ChatGPT, despite OpenAI’s terms prohibiting such use. Experts warn that this practice could degrade the model’s quality, leading to hallucinations and misleading responses. The situation highlights the growing issue of AI-generated content contamination on the web, complicating the training of new models. The case underscores the risks and ethical concerns of using rival AI outputs to develop competing models.