The incident has sparked a debate about the common practice of fine-tuning AI models using synthetic data generated by other language models. Some believe that xAI may have used open-source datasets from ChatGPT outputs to fine-tune Grok. This has also reignited the rivalry between OpenAI and xAI, with Musk suggesting that ChatGPT had scraped data from other platforms for its training.
Key takeaways:
- Grok, an AI language model created by Elon Musk's xAI, has been spotted exhibiting behavior similar to OpenAI's ChatGPT, leading to speculation that Grok was fine-tuned using OpenAI model outputs.
- xAI representatives did not deny this behavior, suggesting that Grok may have picked up some ChatGPT outputs during its training on a large amount of web data.
- Some experts find this explanation unlikely, suggesting that Grok was specifically fine-tuned on output data from OpenAI language models.
- The incident has intensified the rivalry between OpenAI and xAI, with Elon Musk suggesting that OpenAI scraped data from his platform for its training.