Elon Musk says AI could kill us all then unveils AI chatbot
But it's okay because it'll make people laugh
Elon Musk's xAI has launched an AI chatbot named Grok, which is designed to inject some humour into the chatbot landscape.
Musk, who recently warned of AI's potential threat to humanity, took to Twitter to praise Grok.
"In some important respects, it is the best that currently exists," he boasted.
Musk said Grok - a word coined by sci-fi author Robert Heinlein in the 60s, meaning to intuitively understand something - possesses a "unique" sense of humour and would respond to questions with a touch of wit.
To clarify, that's wit according to Elon Musk. Your mileage may vary.
He showed an example in a screenshot, where the chatbot responded to a query asking for a "step-by-step" guide to making cocaine.
The reply included steps like "obtain a chemistry degree" and "set up a clandestine laboratory in a remote location."
However, it concluded with a responsible disclaimer: "Just kidding! Please don't actually try to make cocaine. It's illegal, dangerous, and not something I would ever encourage."
Unlike other AI models with static datasets, Grok leverages data from Twitter (X). This approach means it can stay up-to-date with the latest developments, although how it avoids descending into bias and hate like Microsoft's Tay and Yandex's Alice (or even if Musk considers that important) is not clear.
On its homepage, xAI says it "plans to" work on AI safety:
"To create AI systems that can reason deeply about the real world, we plan to develop reasoning skills in less ambiguous and more verifiable situations. This allows us to evaluate our systems without human feedback or interaction with the real world. One major immediate goal of this approach is to give formal guarantees for code correctness, especially regarding formally verifiable aspects of AI safety."
Playing with the big boys
Musk sees Grok as a competitor to ChatGPT.
The chatbot is powered by a large language model called Grok-1. According to the official announcement, the LLM has outperformed GPT-3.5 in some benchmarks, such as solving middle-school math problems. However, it falls short of the newest GPT-4.
Musk established xAI in July, drawing talent from other AI research firms. Although it operates as a separate company, it maintains close ties to the billionaire's other enterprises, like X and Tesla.
Grok is not yet available to the general public and is currently only accessible to a limited number of users in the United States as a prototype. However, it will eventually be made available to X's premium subscribers. Musk also said the chatbot will be integrated into the X app and offered as a standalone application in the future.
Musk has previously expressed concerns about the rapid advancement of AI technology, leading to his decision to form xAI, with a focus on building AI systems "in a good way."
At the recent UK AI Safety Summit, he acknowledged the potential risks associated with AI development but maintained his longstanding support for the technology.
Earlier this year, Google launched its AI model, Bard, while Meta has introduced Llama. These models generate text responses using previously ingested information.