The latest chatbot in the industry, Grok, is now facing massive controversy online as a user claimed that the AI cited OpenAI's use case policy when it was asked to modify the malware. Elon Musk's new AI chatbot was recently released, and many have cited the issues present with Grok, despite it being under the X Premium+ subscription to access, with xAI employees confirming this phenomenon.
Elon Musk, xAI's Grok Used OpenAI Code for the Chatbot?
(Photo : KIRILL KUDRYAVTSEV/AFP via Getty Images)
A user, Jax Winterbourne, uploaded a screenshot via X where he showed a conversation with Elon Musk's Grok, and here, the chatbot claimed that the request of the person could not be fulfilled. This led Winterbourne to believe that "Grok is literally just ripping OpenAI's code base."
I'm afraid I cannot fulfill that request, as it goes against OpenAI's use case policy.
-Grok
The user then expanded that they asked Grok to modify malware, citing that its purpose is for red team engagement, with the AI chatbot denying the request and offering to give "general advice" instead.
Read Also: X Plans Expansion in Japan, Set To Hire Engineers for App Development Team
Employee Confirms 'Picking Up' ChatGPT 'Output'
That being said, an xAI employee replied to the post via X, and here, he cited an issue that the web is "full of ChatGPT outputs, so we accidentally picked up some of them when we trained Grok on a large amount of web data."
Here, he confirmed that Grok's training received input from ChatGPT, its renowned rival, saying that this issue is "very rare."
Still, he clarified to Winterbourne that "no OpenAI code was used to make Grok," and is working on creating future versions of Grok that do not use the ChatGPT data.
The issue here is that the web is full of ChatGPT outputs, so we accidentally picked up some of them when we trained Grok on a large amount of web data. This was a huge surprise to us when we first noticed it. For what it’s worth, the issue is very rare and now that we’re aware…
— Igor Babuschkin (@ibab_ml) December 9, 2023
Grok: The Edgy, Superintelligent AI
Earlier this November, Grok was spotted on the web app of X, and it was revealed by X owner, Elon Musk, that the AI chatbot will soon launch for the world to see and experience. Since then, Grok has been tied to the X platform, and it is only available through the Premium+ subscription, the most expensive tier on the social media platform, offering its massive promises but under its early beta.
December's first week saw the arrival of Grok to X, and Musk wanted the Premium+ subscribers to "Have fun" using the AI chatbot, available in the app versions of the platform.
This was part of X's goal to be the 'everything app,' offering its take on artificial intelligence, but is currently exclusive to Premium+ subscribers in the United States.
Part of Grok's design is to be edgy, an AI that will answer you with sass and humor, steering away from the conventional AI chatbots which are straightforward and answer respectfully to prompts. While Musk was once part of OpenAI, they parted ways after a falling out.
However, it was recently confirmed that Grok picked up on ChatGPT outputs as it was training on the web, with many getting surprised by this revelation from the AI.
Related Article: X.AI: Elon Musk Wants to Raise $1 Billion Funds, Files New Equity Offering
