Note: This blog post is designed to provide some background regarding my upcoming talk on “belief contracts” at KCON XVI this weekend.
Fake or fact? What is the best way of dealing with false information on the Internet, such as conspiracy theories, fake news, etc.? I recently proposed the creation of a truth market where people could buy and sell belief contracts regarding the truth values of conspiracy theories, disputed media reports (“fake news”), urban legends, and the like. Although I attempted to anticipate and address several possible objections to my proposal when I wrote it up (see especially pp. 8 to 14 of my “truth market” paper), two anonymous reviewers read my work and provided some additional excellent feedback after I submitted my manuscript to the Journal of Free Speech Law. Below is a summary of their main concerns and criticisms:
To begin with, one concern was logistical in nature: how would my truth market work in practice? To the point, how many contracts are initially created when a statement is proposed? How are new belief contracts created? And what is the market cap? Another concern was the possibility of market manipulation, or the potential for “insider trading” by influential media figures. But the main objection to my truth market proposal–an objection made by both of my reviewers–was the open-endedness of my truth market, i.e. its complete lack of a “disciplining function” or “external validation device” (to quote my reviewers).
In brief, unlike a traditional prediction market, where participants buy and sell carefully-worded “event contracts” and have to agree ahead of time to an arbiter to decide whether the event being bet on has occurred or not by a specified date, my truth market would eschew external arbiters and would remain open indefinitely. But without an arbiter or “some trusted person” (to quote one of my reviewers) to “resolve” on some certain date the belief contracts being bet on (i.e. to decide whether a given conspiracy theory or item of fake news is true or not), there would be no foolproof way of assessing whether my truth market was working or not, i.e. whether the price of a belief contract reflected the underlying truth or truth-value of the proposition being bet on.
These are legitimate concerns and objections, and I will respond to them next week.
Pingback: Defending the truth market, part 1 | prior probability