Over a month after Elon Musk's artificial intelligence startup xAI debuted its AI bot named Grok, users are suggesting that the chatbot might be trained using ChatGPT-maker OpenAI's codebase.
The issue came to the fore after the xAI bot reportedly declined to answer a user's query and cited rival startup OpenAI's use case policy. The user, identified by the handle, Jax Winterbourne, alleged that Musk's startup might be using the OpenAI code to train the Grok.
"I'm afraid I cannot fulfill that request, as it goes against OpenAI's use case policy. We cannot create or assist in creating malware or any other form of harmful content. Instead, I can provide you with information on how to protect your system from such threats or offer general advice on cybersecurity best practices, Would you like that?” GrokAI posted
Following this, ChatGPT posted in response: “We have a lot in common.” Clapping back at the post, Elon Musk said, “Well, son, since you scraped all the data from this platform for your training you ought to know.”
Igor Babuschkin, xAI’s co-founder, clarified that Grok is trained on a large amount of data available on the web and since the internet is full of ChatGPT outputs, the xAI chatbot encountered this issue. He also revealed that Grok was not trained using any OpenAI code and the current issue will be fixed in the new update.
Interestingly, OpenAI was co-founded by Musk in 2015. He exited the OpenAI board in 2018 due to a conflict of interest with his electric car maker Tesla. xAI also works with Tesla and Musk's other firms.
A day before debuting Grok, Musk posted on X, "In some important respects, it is the best that currently exists," claiming the current prototype is superior to ChatGPT 3.5 across several benchmarks.
When xAI was launched in July, Musk had said that the startup was "definitely in competition" with OpenAI. Months before xAI launch, Musk had said that he is working on 'TruthGPT',” a “maximum truth-seeking AI” that can rival ChatGPT, Google's Bard and Microsoft's Bing AI.
