David Sacks Claims DeepSeek Used OpenAI Models for Training

Allegations of Knowledge Distillation Against DeepSeek

David Sacks, appointed by Donald Trump as an advisor on AI and cryptocurrency, asserted in a Fox News interview on Tuesday the existence of “substantial evidence” indicating that DeepSeek, a Chinese AI firm, engaged in the unauthorized replication of knowledge from OpenAI’s AI models.

Sacks characterized this alleged action as a form of intellectual property theft. He posited that DeepSeek leveraged responses generated by OpenAI’s models to facilitate the training of its own systems.

Growing Scrutiny of DeepSeek

DeepSeek recently gained prominence due to the release of its high-performing, publicly accessible models and complimentary chat applications. This rapid ascent has prompted heightened examination from U.S. governmental bodies.

Reports from Reuters indicate that the National Security Council is currently assessing the potential ramifications of DeepSeek’s applications, which achieved top rankings within the Apple App Store over the past weekend.

U.S. Navy Restrictions

Concerns regarding security and ethical considerations have led to a reported ban on the utilization of DeepSeek’s AI technology by the U.S. Navy.

This decision underscores the increasing vigilance surrounding the development and deployment of AI technologies, particularly those originating from foreign entities.

Sacks further indicated that OpenAI is likely displeased with the situation, though he did not disclose the specific origins of the “evidence” supporting his claims.