Google and OpenAI complain about distillation attacks that clone their AI models on the cheap

Google and OpenAI are complaining about data theft—yes, you read that right. According to Google, Gemini was hit with a massive cloning attempt through distillation, with a single campaign firing over 100,000 requests at the model, NBC News reports. Google calls it intellectual property theft, pointing to companies and researchers chasing a competitive edge.

Meanwhile, OpenAI has sent a memo to the US Congress accusing DeepSeek of using disguised methods to copy American AI models. The memo also flags China's energy buildout, ten times the new electricity capacity the US added by 2025, and confirms ChatGPT is growing at around ten percent per month.

Distillation floods a model with targeted prompts to extract its internal logic, especially its "reasoning steps," then uses that knowledge to build a cheaper clone, potentially skipping billions in training costs. Google security head John Hultquist warns smaller companies running their own AI models face the same risk, particularly if those models were trained on sensitive business data.