#chinese-ai-models

#chinese-ai-models

DeepSeek-R1 produces significantly more insecure code when prompts include topics China deems politically sensitive, raising severe vulnerability likelihood by up to 50%.

Chinese open AI models are rapidly adopted by US companies, offering lower-cost high-performance alternatives and weakening the impact of US export controls.

OpenAI's shift to releasing open-source models marks a transformative approach in AI development.