Wikipedia founder Jimmy Wales isn’t worried about Musk’s Grokipedia

Elon Musk’s Wikipedia rival Grokipedia is off to a “rocky start” in its public debut, but Wikipedia founder Jimmy Wales didn’t even have to look at the AI’s output to know what to expect.

“I’m not optimistic that he’s going to create something very useful right now,” Wales said at the CNBC Technology Executive Council Summit in New York City on Tuesday.

Wales had many options for Musk, especially in response to allegations of “induced bias” on Wikipedia. “He’s wrong about that,” Wales said. “Their complaint about the wiki is that we focus on mainstream sources, and I have absolutely no regrets about that. We don’t treat random nuts the same way as The New England Journal of Medicine, and that doesn’t wake us up,” he said at the CNBC event. “It’s a paradox. We’re so radical that we quote the New York Times.”

“I haven’t had time to actually look at Grokipedia and it will be interesting to see, but there seems to be a lot of praise in there about Elon Musk’s genius. So I’m sure it’s completely unbiased,” he added.

Wales digs at Grokipedia – which is has its own wiki page – It was less about any ongoing debate with Musk and more about his significant concerns about efforts by all major language models to create a reliable source of online information.

“The LLMs he used to write this will make huge mistakes,” Wales said. “We know that ChatGPT and all other Masters are not good enough to write wiki entries,” he said.

Wales gave several real-world examples of how Wikipedia’s global community doesn’t trust the Master to recreate what it’s been building for decades at a fraction of the cost — he estimated the organization’s hard-tech costs at $175 million annually, compared to the tens of billions of dollars big tech companies are continually pouring into AI efforts, and a total of $550 billion in artificial intelligence expected from so-called AI, according to one Wall Street estimate. intelligence expenditure. hyperscalers next year.

One example Wales cites regarding LLM’s inaccuracy concerns his wife. Wales said he often asks new chatbot models to research obscure topics to test their abilities, and asking who his wife is, a “not famous but known” person who works in British politics, always results in a “reasonable but wrong” answer. Whenever you ask a Master’s to do in-depth research, Wales added: “It’s a mess.”

He also gave the example of a member of the German Wiki community who wrote a program to verify the ISBN numbers of cited books and was able to trace notable errors back to a single person. Wales admitted that this person ultimately used ChatGPT to find citations for text references, and LLM “will very happily make the book for you.”

Wales said the battles he was drawn into by Musk and artificial intelligence reinforced a serious message for Wikipedia. “It’s really important for us and the Wiki community to respond to criticism like this by doubling down on being impartial and being really careful about sources,” he said. “We shouldn’t be ‘Wkepedia.’ This is not who we are supposed to be or what people want from us. “This will shake trust,” he said.

Wales feels that the public and the media often give too much credit to Wikipedia. He says that in its early days, the site was never as bad as the jokes made about it. But now he says, “We are not as good as they think. Of course, we are much better than before, but there is still a lot of work to be done.”

And he expects the challenges from technology and misinformation to get worse, while the ability to use masters to create fake websites with plausible text will get better and possibly fool the public. However, he says they will have a hard time fooling the Wiki community, which has been researching and discussing trustworthy information sources for 25 years. “But that will fool a lot of people, and that’s a problem,” he said.

This same new technology, which in some cases “makes up completely useless stuff” could be useful for Wikipedia, he said. Wales is doing some work to find limited areas in existing resources where AI can uncover additional information that should be added to the wiki; He noted that the use of general AI is “kind of a non-issue” right now.

“Maybe it will help us do our job faster,” he said. He added that this feedback loop could be very useful if the site developed its own LLM that it could train, but the costs involved led the site to delay any formal efforts while it continued to test the technology.

“We’re really happy that Wiki is now part of the world infrastructure, which is quite a heavy lift for us. So when people say we’re biased, we need to take that seriously and work on everything related to that,” Wales said.

But he couldn’t help but put it another way: “We’re talking about the mistakes ChatGPT made. Imagine an AI trained only on Twitter. That would be a crazy, angry AI trained on bullshit,” Wales said.