Elon Musk testifies that xAI trained Grok on OpenAI models

2026-04-30 · Source: TechCrunch · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Novice, quick

Summary

OpenAI and Anthropic are actively combating "distillation," a process where third parties train new AI models by prompting their publicly accessible chatbots and APIs. This technique allows other firms, including Chinese companies, to create open-weight models nearly as capable as U.S. offerings but at a significantly lower cost. Elon Musk recently admitted in a California federal court that xAI has used distillation on OpenAI models to train Grok, characterizing it as a general practice among AI companies. This admission highlights how distillation threatens the competitive advantage of AI giants built on extensive compute infrastructure. OpenAI, Anthropic, and Google are reportedly collaborating through the Frontier Model Forum to share information and prevent suspicious mass queries aimed at distillation.

Key takeaway

For CTOs and VPs of Engineering evaluating AI model development strategies, Elon Musk's admission confirms that distillation is a prevalent, albeit controversial, method for rapidly developing competitive models like Grok. You should assess your organization's exposure to distillation, both as a potential target and as a development tactic, while carefully reviewing API terms of service to mitigate legal and competitive risks.

Key insights

AI model distillation, a method of training new models from existing APIs, is a widespread industry practice.

Principles

• Distillation undermines compute infrastructure advantage.
• Terms of service may prohibit distillation.

Method

Distillation involves systematically querying existing models to understand their inner workings, enabling the creation of new, similarly capable models at lower cost.

In practice

• Monitor API query patterns for anomalies.
• Implement rate limits on mass queries.

Topics

• AI Model Distillation
• xAI Grok
• OpenAI Lawsuit
• Elon Musk Testimony
• Frontier Model Forum

Best for: CTO, VP of Engineering/Data, Executive, Legal Professional, Tech Journalist, Director of AI/ML

Related on AIssential

• AI Distillation Explained: The Truth Behind the Biggest AI Controversy Right Now — AI model distillation attacks are a significant and escalating threat to intellectual property and frontier model capabilities.
• Who's Really Stealing From Whom? — Blackbox distillation allows smaller models to learn from frontier models' API outputs, even without access to internal probabilities.
• Musk v. Altman week 1: Elon Musk says he was duped, warns AI could kill us all, and admits that xAI distills OpenAI’s models — The Musk v. OpenAI trial centers on alleged deception regarding OpenAI's mission and the stewardship of AI safety.
• Mustafa Suleyman's case against open-source AI shortcuts — Distillation limits open-source AI's general utility, creating a significant performance gap with frontier models.
• How much does distillation really matter for Chinese LLMs? — Distillation, while a legitimate training method, is being used illicitly by Chinese labs to extract capabilities from US frontier AI models.
• Musk confirms: xAI to open-source Grok 3 — xAI is open-sourcing Grok 3, aligning with its strategy to release prior-generation models and enhance transparency amid regulatory scrutiny.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by TechCrunch.