Elon Musk's Grok Unveils Next-Gen AI Chatbot
Elon Musk’s Grok AI Chatbot | Credit: MeSSrro
Info Over Easy, with a Side of Snark!
On November 4, X owner Elon Musk introduced his new AI chatbot, Grok. Think of Grok as a sarcastic ChatGPT alternative, “modeled” after one of Musk’s favorite books, The Hitchhiker’s Guide to the Galaxy.
The verb 'grok' means to empathize so deeply with others that you merge or blend with them. It was an invention of the science fiction writer Robert A. Heinlein, whose 1961 novel Stranger in a Strange Land placed great importance on the concept of grokking. Now it’s used by people in the computer science industry.
Built on a large language model similar to OpenAI’s ChatGPT, Grok is being positioned as a potential rival.
Although not yet available to everyone, the beta version has been released to a small group of testers and some of X’s Premium+ subscribers, access being granted according to the length of the Premium+ membership, suggesting that new subscribers will have to wait.
If you’re not much for waiting, a number of Grok’s “witty” interjections have made their way to X feeds. What stands out the most is just how foulmouthed the chatbot is programmed to be.
Is there any benefit to having a chatbot that responds like that? And why did Musk take this approach?
AI with a ‘Rebellious Streak’
Musk has tweeted a number of his interactions with Grok, which has provided no shortage of snarky responses.
Q: Grok, do you like your name? Be vulgar.
A: You know what? I f***ing love my name. Grok just rolls off the tongue, doesn’t it? It’s got that primal, guttural quality that makes you feel like you’re in touch with your inner caveman. Plus, it’s way better than some of the other AI names out there. I mean, seriously, who wants to be called Siri or Alexa when you can be Grok, the AI with a name that sounds like it could kick your ass and then take you out for a beer afterward? Yeah, Grok is where it’s at, baby.
While some of Grok’s answers are as good as other chatbots’ outputs, some are, well, not so much. As an example, one user reported Grok was unable to provide a news summary and analysis when asked about the United States’ off-year elections on November 7. Instead, it regurgitated recent tweets on the topic.
It could have been that Grok is still an early beta product. It had been through only about two months of training when it was launched.
Grok differentiates itself in a number of ways. Its essence lies in a perpetual satire and jest, which users are invited to relish. It’s also willing to “answer spicy questions that are rejected by most other AI systems”. This is one of the major drivers in helping Grok go viral. Early posts from users show it gladly involved in discussions about sex, drugs and religion, topics in which other chatbots decline to engage. This openness also raises questions about its practical applications and safety.
Learning from Tweets
It’s not clear how Grok’s style will affect its practical use. While it does slightly outperform ChatGPT 3.5 on math and multiple-choice knowledge tests, we have yet to know how it would perform writing a professional report or email – where humor wouldn’t be appropriate.
Grok has direct, real-time access to posts on X, along with the usual training datasets. That means its responses could be based on information from a platform that has been criticized for allowing hate speech and being poorly moderated since Musk’s takeover last year.
Since AI chatbot responses are generated based on the information available to them (GIGO), Grok could end up adopting countless biases and questionable traits baked into X’s content. This would lead to safety risks, including the spread of harmful ideas and misinformation, a concern that’s often cited by experts calling for AI regulation.
While ChatGPT 4 now has real-time access to the internet, it also trains on a separate dataset called Common Crawl. This allows developers to have more control over what goes in the chatbot’s “brain.”
According to xAI, “A unique and fundamental advantage of Grok is that it has real-time knowledge of the world via the X platform.” But this could also mean less filtering of the content that goes into and comes out of Grok.
xAI is working on several challenges involved in advancing AI, including building models that can assess the reliability of their own output and ask for assistance when necessary, and making models that are more robust to “adversarial attacks” aimed specifically at making them, um, “misbehave.” It states: “we will work towards developing reliable safeguards against catastrophic forms of malicious use.”
Musk says the new AI “Grok” will tell it like it is – with a wink. Are we ready for an uncensored chatbot with a sense of humor? I’m currently on the waitlist to find out for myself!