Elon Musk has publicly admitted that his AI venture xAI used models from OpenAI to help train its chatbot Grok. The statement came during legal proceedings linked to Musk’s ongoing dispute with OpenAI.
According to reports, Musk acknowledged using ‘model distillation,’ a controversial industry practice that is rarely disclosed publicly.
Model distillation is a technique in which a larger, more advanced AI system is used to train a smaller model, transferring knowledge and improving performance. The approach is widely used within organizations to optimize their own models. However, it has also become a point of debate when applied across competing companies.
Tech giants such as Anthropic and Google have previously warned that distillation can replicate capabilities without equivalent investment, raising uncomfortable questions about who owns what in AI development.
Musk's admission that xAI used OpenAI's models has raised further questions about his relationship with the company. It also hands Altman a concrete example to point to in his long-running dispute with Musk.
Some experts argue that distillation lowers barriers to entry for new competitors. The process threatens to undermine the competitive advantage of companies. The debate now centers on two paths: regulating the practice or accepting model distillation as an industry standard.
Looking ahead, there will be a need for clearer regulatory definitions, placing pressure on both policymakers and companies. The current situation demonstrates the interconnection between different areas of AI research. The progress of each organization affects the ability of competing companies to succeed.
Also read: Elon Musk Challenges Sam Altman & OpenAI’s Business Ethics as Court Trial Begins