Microsoft and OpenAI launch an investigation to probe DeepSeek data access of compromised information using OpenAI's API for model building.

The joint investigation by Microsoft and OpenAI seeks to establish facts regarding unauthorized access of OpenAI model data by groups associated with Chinese AI startup DeepSeek. In the autumn Microsoft security experts detected abnormal activities as potential DeepSeek employees extracted substantial amounts of data through OpenAI’s API. DeepSeek hesitated to meet OpenAI's data accessibility parameters which made companies deeply concerned.

DeepSeek's newly released open-source AI model named R1 has started a competition through benchmark tests where its technology demonstrates better performance than US models for mathematical reasoning alongside general knowledge. The startup potentially utilizes OpenAI-made data to strengthen its artificial intelligence advancement leading organizations to sound alarm bells.

As the former leader of Trump's team of AI experts David Sacks claimed that DeepSeek used "distillation techniques from OpenAI models" to train their systems using OpenAI's gained knowledge. OpenAI confirmed ongoing attempts made by foreign entities targeting US-based AI models for knowledge extraction yet refused to discuss the DeepSeek situation directly.

An inquiry by Microsoft and OpenAI confirms rising concerns regarding protecting data during worldwide AI competition.