Grok, Elon Musk’s recently launched chatbot, is facing scrutiny as users have noted similarities between its responses and those generated by OpenAI’s ChatGPT. Observations include instances where Grok explicitly cites OpenAI in its replies. Users, such as Jax Winterbourne and Pezzo, have shared these findings, raising concerns about potential code replication.
Notably available to X Premium Plus subscribers, Grok has drawn attention for what some perceive as content extraction from ChatGPT. Pezzo shared a response from Grok instructing users to report issues to “support@openai.com,” further fueling suspicions.
In response to the user observations, xAI engineer Igor Babuschkin admitted that during Grok’s training, data from ChatGPT was inadvertently incorporated. He clarified that this issue is rare and assured users that future versions of Grok would address and eliminate such occurrences. Babuschkin emphasised that no OpenAI code was used in the development of Grok.
Elon Musk, upon Grok’s November launch, touted it as a leading AI model with real-time access to information on X and internet browsing capabilities akin to ChatGPT. Musk highlighted Grok’s conversational question-answering abilities, coupled with a touch of humor in its responses.
This incident follows Musk’s previous criticism of AI firms, including OpenAI, for aggressive scraping of Twitter data to train large language models. Musk raised concerns about the impact on user experience due to widespread data scraping practices associated with generative AI models.
