June 3, 2025:
DeepSeek Suspected of Using Googles Gemini Data - Chinese AI lab DeepSeek is speculated to have used Google's Gemini model outputs to train its latest AI model, R1-0528. Evidence shows similarities in language and traces between DeepSeek's and Gemini's models, but conclusive proof is missing. DeepSeek has been accused before of exploiting rival data, including OpenAI's ChatGPT.
The AI industry is struggling to secure model outputs. OpenAI and other companies are implementing stricter access measures to prevent unauthorized training and distillation practices. This comes amid growing concerns about data contamination from AI-generated content.
