Elon Musk just confirmed what many suspected. During testimony, he acknowledged that xAI used OpenAI's models to help train Grok, his competing AI chatbot.
This matters because it touches on one of the hottest controversies in AI right now: distillation. That's when you use a powerful model's outputs to train a smaller, cheaper one. It's technically legal in most cases, but frontier labs like OpenAI, Anthropic, and Google are increasingly trying to ban it in their terms of service.
The irony here is thick. Musk co-founded OpenAI, left after disagreements about its direction, then sued the company claiming it betrayed its open-source mission. Now he's admitting his own company used OpenAI's work as training data.
For anyone building with AI, this highlights a growing tension. The big labs want to protect their competitive moats by preventing distillation. But if you're training models or fine-tuning systems, understanding what's allowed and what violates terms of service is becoming critical.
The legal landscape is still murky. Courts haven't definitively ruled on whether distillation infringes on model creators' rights. Musk's testimony might become a reference point as these cases move forward.
Bottom line: if you're using AI model outputs to train or improve your own systems, pay close attention to the terms of service. What's technically possible and what's legally permitted are two different things, and the rules are tightening.
