Decentralisation will eliminate bias, and will allow fact-checking the fact-checkers.
Who is spreading “fake news”? What should be done about it?
When Li Wenliang tried to sound the alarm on COVID-19, the Chinese government summoned him for “making false comments on the Internet”. They declared he was fake news. It really can’t get any more obvious how bad of a system that is. To decide by decree what is true and what is false is an extremely dangerous authoritarian idea that belongs only in Orwell’s 1984. But even in the West, too many of the proposed solutions to “fake news” are about limiting peoples ability to share information. This is not what people want. People want access to the facts, and then to form their own opinion, based on those facts. Nobody wants to be told what to think or what they can’t share.
Facebook, Google, and Twitter all have rules about what information is allowed and what is not. They justify some of these arbitrary rules as preventing misinformation. Twitter says you cannot go against guidance from global health authorities. Youtube has decided the true origins of the virus, and any dissenting opinion is banned. Facebook did the same, and so did Twitter.
There is a hermetically sealed bubble of elite academe, prestige liberal media and Silicon Valley nouveaux riches, impenetrably convinced of their absolute moral and intellectual rectitude. And they want to control — and do control — what you can read and think.— Sohrab Ahmari (@SohrabAhmari) April 18, 2020
Everyone is just calling each other fake news with complete disregard for the facts. If people are putting out false information, then its our responsibility to put out true information - to provide the tools and information necessary for people to make more informed decisions.
Fake news leads to public ignorance, ill-informed decisions, and leads to further distrust in the media as a whole which also limits the ability of good journalists to do their job. Telling people what to think, or restricting what news they’re allowed to share, also leads to these same outcomes. I want to live in a society where people are free to express themselves and have a voice — not always trying to shut each other down and deplatform their enemies for political reasons.
A significant danger of fake news is when the mainstream and credentialed media gets it wrong. And their own fact-checkers aren’t going to say anything about that. When The New York times published fake news claiming Iraq had weapons of mass destruction, it was used as a false premise to launch a war. Thousands of innocent people died. Had we known the truth, those people could still be alive today. Perhaps we could have responded to COVID-19 faster instead of saying it was just the flu.
Existing fact checking organisations (such as Snopes.com and PolitiFact.com) can be effective, but they have limits such as when there is a perception of partisan bias. During the 2016 US election the Tampa Bay Times made political endorsements and this deligitimised the non-partisanship of its PolitiFact.com fact-checker. It’s now owned by a different entity, but it’s still viewed as having that bias. Even if centralised fact checkers only publish the truth, they can still be subject to selection bias by preferentially fact-checking one side. When 1,000 American voters were polled on whether they think media fact-checkers are biased, 62% said they skew the facts to help candidates they support.
To stop fake news, we need lots of people to be fact-checking. We need lots of people to be skeptical. Before you share that article saying Russia is releasing 800 lions onto the streets to enforce a lockdown, just take a few seconds to look into the source or do a quick search. It takes a widespread effort to prevent falsehoods spreading, and leaving the job of fact-checking to a handful of media organisations isn’t enough.
So this is where Verifact comes in, as a decentralised fact-checker.
Anyone can submit articles or tweets to the Verifact platform for fact-checking. This solves the selection bias problem.
Verifact lets users vote on the factuality by staking tokens towards the verdict they deem true. If you stake tokens on the verdict that is deemed factual by the network, you will gain more tokens from those who voted opposite to you. This creates an incentive for authentic fact checks.
A decentralised fact-checker is not straight-forward though. The naive system of just having upvotes and downvotes on whether something is true won’t work, and will just create an echo chamber. We don’t want to create a popularity contest because that just scares journalists into not telling the truth when it’s unpopular — one of the most important parts of their job. A decentralised fact-checker needs to have sensible incentives, and needs to be self-correcting in its search for the truth. The Verifact protocol does this by allowing anyone with the truth to stake against the entire crowd, if they are willing to front the tokens to do it. Rather than rewarding agreement with the crowd, it rewards a minority opinion that sways the crowd.
Verifact will use a game-theoretic structure with similar incentives to a prediction market, by using Robin Hanson’s logarithmic market scoring rule market maker (LMSR). This allows people to profit from movements in opinions, which is a better system than whether the crowd agrees at that point in time. The whitepaper is still being written, and a draft will be published in the coming days. Sign up at https://verifact.io to stay updated.
The incentives of Verifact are designed to eliminate bias — to incentivise honest fact-checks from people with differing opinions, to build a consensus on what is fake news and what is true, and why.
I’m interested to hear what you think about this. Please complete the survey at https://verifact.typeform.com/to/QdOsnD
Thanks for reading :)
P.S. Any ideas? Please send me a message on Twitter.