DeepSeek, a Chinese AI research lab originating from High-Flyer hedge fund, has released DeepSeek-R1, an advanced AI model rivaling OpenAI's capabilities in mathematical reasoning and code generation while using fewer resources.

The model, released under an MIT license with six smaller variants, employs reinforcement learning and multi-stage training. This has prompted OpenAI and Microsoft to investigate whether DeepSeek trained their model using OpenAI's outputs, potentially violating terms of service.

David Sacks, dubbed the 'AI Czar,' suggests there is "substantial evidence" that DeepSeek used knowledge distillation - a technique where one model learns from another by asking millions of questions to mimic its reasoning process.

OpenAI CEO Sam Altman expressed concern about copying existing work versus innovative development, stating: "It is (relatively) easy to copy something that you know works. It is extremely hard to do something new, risky, and difficult when you don't know if it will work."

The situation highlights a paradox in AI development: OpenAI defends its practice of scraping internet data for training, including from potentially illegal sources, while opposing knowledge distillation between models. An OpenAI spokesperson emphasized their commitment to protecting intellectual property and working with the US government to safeguard advanced AI capabilities from competitors.

The controversy raises important questions about AI development ethics, intellectual property rights, and the global competition in artificial intelligence advancement.

Related Articles

Previous Articles