Vitalik Buterin Says Grok Keeps Musk’s X Honest

X’s artificial intelligence chatbot, Grok, makes the social media platform more truth-friendly with responses that oppose users who turn to it to confirm political biases, Ethereum co-founder Vitalik Buterin said.

“The easy ability to call Grok on Twitter is probably the biggest thing after community notes that has been positive for the truth-friendliness of this platform,” Buterin on Thursday. 

”The fact that you don’t see ahead of time how Grok will respond is key here,” he added. “I’ve seen many situations where someone calls on Grok expecting their crazy political belief to be confirmed and Grok comes along and rugs them.”

Buterin a strong case can be made that Grok is a “net improvement” to X, but acknowledged concerns over how the AI chatbot is fine-tuned to learn from the opinions and views of certain users, including its creator, Elon Musk.

Grok’s flaws were on full display last month when it and even suggested he could have resurrected faster than Jesus Christ.

Musk “adversarial prompting” for the hallucinations, and crypto executives argued it showed why AI needs to be decentralized in order to , and impartiality.

AI asserting views as fact a real concern

Kyle Okamoto, chief technology officer at decentralized cloud platform Aethir, told Cointelegraph last month that “when the most powerful AI systems are owned, trained and governed by a single company, you create conditions for algorithmic bias to become institutionalized knowledge.” 

“Models begin to produce worldviews, priorities and responses as if they’re objective facts, and that’s when bias stops being a bug and becomes the operating logic of the system that’s replicated at scale.”

Related:

Grok is built by Musk’s AI company, xAI, and is one of the most widely used . With over 1 billion people using AI, incorrect and misleading information has the potential to spread rapidly.

Still, Buterin said Grok has been more successful in making X more truth-seeking than many of the other “third-party slop that we see.”

AI chatbots need to improve across the board

Grok isn’t the only with problems — OpenAI’s ChatGPT has been criticized for biased responses and factual errors, while Character.ai is that its chatbot lured a 13-year-old boy into a sexually abusive interaction and even encouraged him to take his own life.

Magazine: