Elon Musk Seemingly Admits xAI Has Used OpenAI’s Models to Train Its Own

Editor's take

Elon Musk's deposition testimony suggests xAI has leveraged elements from OpenAI's models during its own training, a practice he frames as common industry behavior.

This admission is significant because it raises questions about xAI's claims of independent innovation and could impact the competitive landscape, especially given OpenAI's previous concerns about data scraping and Musk's departure from its board. The implication is that xAI's progress, potentially including its Grok model, may not be as purely original as initially presented, complicating its positioning against established players like Google's Gemini and Meta's Llama.

Future scrutiny will likely focus on the extent of this cross-training and whether it constitutes a violation of any existing agreements or industry norms. The reaction from OpenAI and other AI developers, alongside any potential legal ramifications or industry-wide policy shifts, will be crucial to observe.