Elon Musk reveals Grok, a rebellious AI chatbot

Elon Musk reveals Grok, a rebellious AI chatbot

Elon Musk has unveiled his latest creation in the world of artificial intelligence – an AI chatbot named Grok with a “rebellious streak” inspired by the beloved sci-fi comedy series The Hitchhiker’s Guide to the Galaxy.

The eccentric billionaire and Tesla CEO, who last week warned a gathering of experts at Bletchley Park that uncontrolled AI could become “one of the biggest risks to the existence of civilization,” gave some initial details about Grok in a series of tweets over the weekend.

Musk shared what appears to be a screenshot of a conversation with Grok, asking it to provide a “step-by-step” guide to making cocaine. Grok obliges by outlining four tongue-in-cheek steps, ending its response with: “Just kidding! Please don’t actually try to make cocaine. It’s illegal, dangerous, and not something I would ever encourage you to do.”

This seeming attempt at humor reflects Grok’s intended personality as an AI assistant inspired by the witty and irreverent Hitchhiker’s Guide, Musk explained. The name Grok itself comes from a word coined by sci-fi author Robert Heinlein, meaning to understand something thoroughly and intuitively.

After limited testing, Musk plans to offer access to Grok as part of the premium subscription plan on X, the social media platform he now owns. This would put Grok in direct competition with ChatGPT, the buzzy AI chatbot from his former company OpenAI.

Grok is the first product launched by Musk’s newly formed AI startup xAI. In a blog post, xAI revealed that Grok is powered by an AI model called Grok-1, which they claim has surpassed the performance of GPT-3.5 that powers ChatGPT’s free version. However, Grok-1 still falls short of ChatGPT’s most advanced proprietary model GPT-4.

By sharing Grok’s tendency for sarcastic quips and warning not to “use it if you hate humor,” xAI seems to be encouraging more casual and experimental use compared to ChatGPT’s aversion to potential harm. This aligns with Musk’s stated goal for xAI to develop AI “in a good way” that remains under human control.

The rollout of Grok represents a change of course for Musk, who previously expressed concerns about AI progress moving too quickly. His new embrace of building advanced models like Grok underscores the rapidly evolving landscape of AI chatbots and their potential to transform how humans interact with machines.