The Seven Rules of Trust: Why It Is Today’s Most Essential Superpower Jimmy Wales Bloomsbury (2025)
Jimmy Wales, an Internet entrepreneur from Huntsville, Alabama, now based in London, is best known for creating Wikipedia, which launched in January 2001. The online encyclopedia now holds more than seven million articles and has become a standard guide for anyone seeking information.
The Wikimedia Foundation, a non-profit organization in San Francisco, California, runs the project with about 700 employees, but Wikipedia still relies entirely on unpaid volunteers to write and edit its articles: hundreds of thousands of people contribute to the site each month, under a set of community-developed rules to deal with disagreements, cut down self-promotion and generate consensus.
Wales’ book, The Seven Rules of Trust, released in October, tells the tale of Wikipedia’s origins and how the project highlights lessons in earning and maintaining trust. Wales spoke to Nature about the importance of scientific transparency and the rise of artificial intelligence.
Anyone can contribute to Wikipedia. Does the site lack respect for expertise?
Everything about Wikipedia is a worship of expertise. We have rules about citing quality sources: newspapers, magazines, academic journals, peer-reviewed literature and books from reputable publishers. That’s very different from letting anyone offer a random opinion.
It is true that you don’t have to be an expert to write a Wikipedia article. But, for lots of topics, having amateur passion is a great way to get started. It’s like journalism: journalists have to write about topics that they aren’t experts in, but hopefully they consult and quote the experts. Then, the journalist can make sense of it.
Your seven rules of trust are: be personal; collaborate; have a mission; give trust to get trust; stay civil; don’t take sides; and be transparent. Which one is most relevant to scientists?
‘Be transparent’ is very important. That’s not an admonishment that science needs to become more transparent — it’s an observation that science is very transparent. There are obviously areas in which this could be improved, but I think that the statement is generally true.
There was a study of the ‘retraction penalty’, which found that running a correction for a scientific article is better for an author’s citation counts than is trying to bluff their way through it (S. F. Lu et al. Sci. Rep. 3, 3146; 2013). I thought that was super interesting research. The conclusion makes sense, but it’s nice to see science validate it.
Wikipedia prides itself on presenting balanced points of view. Is that hard?
That will always be a struggle. We think a lot about ‘wikivoice’: when do we say something in the voice of Wikipedia? Declaring ‘this is the fact’ is high risk.
We look for consensus not just in the sources, but also in the community of Wikipedians. If you have Wikipedians in good standing talking and editing in good faith, and no one is willing to stand up and say, ‘actually, I think that the evidence is still mixed’, then you can go ahead and use wikivoice.
It’s really important that we be as intellectually diverse as possible. If you ask an English speaker who invented the aeroplane, they will say the Wright brothers — like that’s a simple fact that we all learnt when we were six years old. But if you ask a French person, they might say something different: a Brazilian inventor [Alberto Santos-Dumont] who was living in France. We try to consider all sides of an issue. If we are not even aware of all sides of the issue, it becomes that much harder.
How do you try to increase the diversity of Wikipedia writers and editors?
We run ‘edit-athons’. Those have been successful in adding women in science as Wikipedians. But it’s hard. We have to be thoughtful about the environment of the site: is it open? Is it welcoming?
I don’t have any statistics; I’m afraid the demographics haven’t changed nearly enough. We talk about it a lot. I want more people from different backgrounds to edit Wikipedia — so, anyone reading this, please come and help us.
You don’t mention ‘accountability’ as a pillar of trust in your book.
The concept is there. Our writers can use pseudonyms; many do. But Wikipedia runs on an accountability culture, rather than a gatekeeping culture. All edits that you make are tied to your account, and people can see your track record and your history. If you are not doing a good job, people will call you out. Accountability is incredibly important.
Do all editors act in good faith?
We do see bad-faith actors who will solicit business by saying ‘I’m a Wikipedia administrator’ (they aren’t), claiming to be able to help write one-sided articles. I think it’s a bigger problem for their victims than it is for Wikipedia, because Wikipedia will require sources, and if the sources aren’t there, then the edits won’t stick.
Do you see more problems now that artificial intelligence bots can write material?
This is something we need to be vigilant about, but it hasn’t yet been a major problem. Sheer volume isn’t really how Wikipedia works. If you start making ten edits a minute, you are blocked immediately. That’s obviously a bot doing something nefarious.
In fact, we don’t ban the use of AI. We do say, be very careful with it, and you’re responsible for what you put in Wikipedia.
A lot of people write articles in a language that isn’t their first, so they use AI to help them with grammar. That sounds great to me — wonderful, more participation.
Do you see generative AI as a threat to trust?
The hallucination problem — when large language models produce responses containing nonsensical or fabricated information — is still very bad. And the more obscure the topic, the worse the hallucinations become.
The sentences produced by chatbots sound very plausible, so I do think that’s a problem for trust. I am hoping that people’s reaction to this will be a renewed interest in journalism, old-school fact-checking and human oversight. If people think that the major news services are just ‘part of a conspiracy’, then we’re lost.
Can AI be useful for trust?
Wikipedia is very transparent, but in some ways it’s so transparent that it’s opaque. If you want to know why a certain article is written a certain way, go to its talk page, and you can read all about it. But, sometimes, that’s 100 pages of debate. Who has time to read 100 pages?
