{"id":39172,"title":"OpenAI Claims Evidence of Chinese AI Firm DeepSeek Utilising Its Model for Training","publisher":"Stockmark.IT","author":"Stockmark.IT Website","published":"2025-01-30T04:40:29+00:00","modified":"2025-01-29T17:40:40+00:00","canonical_url":"https://stockmark.it/openai-claims-evidence-of-chinese-ai-firm-deepseek-utilising-its-model-for-training/","markdown_url":"https://stockmark.it/openai-claims-evidence-of-chinese-ai-firm-deepseek-utilising-its-model-for-training.md","json_url":"https://stockmark.it/openai-claims-evidence-of-chinese-ai-firm-deepseek-utilising-its-model-for-training.json","category":"Artificial intelligence","categories":["Artificial intelligence"],"featured_image":"https://i0.wp.com/stockmark.it/wp-content/uploads/stencil.default-2025-01-29T173956.118.jpg?fit=1200%2C800&quality=89&ssl=1","format":"news","language":"en-GB","content":"In a significant development within the artificial intelligence sector, OpenAI has announced it has discovered evidence suggesting Chinese AI startup DeepSeek employed OpenAI’s proprietary models to train its own open-source competitor, raising serious intellectual property concerns.\n\nThe San Francisco-based ChatGPT creator revealed it had witnessed instances of “distillation”, a technique where developers extract performance capabilities from larger models to enhance smaller ones. Whilst distillation is a standard industry practice, the core issue lies in DeepSeek’s alleged use of this method to construct a rival model, directly violating OpenAI’s service terms.\n\nDeepSeek’s introduction of its R1 reasoning model has sent ripples through the market, particularly impacting tech giant Nvidia, whose shares plummeted 17 per cent on Monday, erasing £589 billion in market value. The dramatic decline stemmed from growing concerns that substantial investments in costly AI hardware might become unnecessary.\n\nSources close to the matter indicate that OpenAI and Microsoft conducted investigations last year into suspected DeepSeek accounts utilising OpenAI’s API. These accounts were subsequently blocked due to suspected violations of service terms through unauthorised distillation practices.\n\nThe situation has drawn attention from high-profile figures, including David Sacks, President Trump’s AI adviser, who acknowledged the possibility of intellectual property theft through distillation techniques. Industry experts note that Chinese AI laboratories commonly utilise outputs from established companies like OpenAI, who have invested heavily in human-aligned response training.\n\nOpenAI faces its own challenges regarding copyright infringement, with ongoing lawsuits from prominent publishers and authors claiming unauthorised use of their content for model training. The company maintains it actively implements countermeasures to protect its intellectual property and collaborates with the US government to safeguard advanced models from competitor exploitation."}