Story perspectives

OpenAI Warns Lawmakers: DeepSeek Threatens U.S. AI Security

2/13/2026

Summary

Story summary

OpenAI has alerted U.S. lawmakers about DeepSeek's efforts to extract outputs from OpenAI models to train competing systems.
In a memo, OpenAI accuses DeepSeek of bypassing access restrictions and using distillation, where a newer model learns from a stronger one.
The practice raises concerns about national security and erodes U.S. AI advantages.
OpenAI notes that DeepSeek's chatbot censors sensitive topics and investigations into its data acquisition have begun.

1 of 1

Story summary

OpenAI has alerted U.S. lawmakers about DeepSeek's efforts to extract outputs from OpenAI models to train competing systems.
In a memo, OpenAI accuses DeepSeek of bypassing access restrictions and using distillation, where a newer model learns from a stronger one.
The practice raises concerns about national security and erodes U.S. AI advantages.
OpenAI notes that DeepSeek's chatbot censors sensitive topics and investigations into its data acquisition have begun.