Thread by @jonathanstray

Meta will stop paying pro fact checkers, and switch to a community notes system. In part, they say this is because of fact checker "bias."

I don't think this is good news. But it must be said that the fact checking community did this to itself.

How?
🧵

about.fb.com/news/2025/01/m…

In my view fact checking has two goals:
1) Identify harmful falsehoods
2) Create trust with the audience

The fact checking community mostly succeeded at #1. And they mostly failed at #2. They were mostly right, but many people no longer believed them.

Here's the evidence...

First, were fact checkers trusted? In the US it was largely split along partisan lines. Rep thought they were biased.

Unfortunately, there was also far more false information circulating on the Rep side. That is also very well documented...
poynter.org/ifcn/2019/most…

No one wants to review the evidence that there was more misinfo on the right than the left. If you're on the left, it seems obvious. If you're on the right, it just sounds biased.
But the evidence is very consistent across studies and methods. E.g.
science.org/doi/10.1126/sc…

Most of these studies use fact checker ratings as ground truth -- so it's easy to say the labels are biased!
Now it gets interesting. When you ask politically balanced groups of regular people to identify misinfo, they mostly agree with the pros!

https://x.com/jonathanstray/status/1709665375703675151

In other words, the fact checking community was both
1) mostly correct in their calls
2) mostly distrusted by the people who were exposed to the most misinfo

And there was an obvious thing they could have done to fix #2. Namely: involve conservatives in the process.

The biggest factor which determines whether fact checkers are trusted or not is whether the fact check is coming from the outgroup.
No trusts their political opponents to play fair when determining truth. Nor should they, honestly!

https://x.com/diegoreinero_/status/1657086253400215554

Unfortunately, the professional fact checking community was essentially an appendage of the professional journalism community.
This meant rigorous standards, but it also meant almost everybody politically left-ish.
Here's American journalists' politics.
niemanlab.org/2024/12/newsro…

In short: there was a simple thing fact checking orgs could have done to build trust: hire conservative fact checkers. This wouldn't even have much changed the calls they made!
There is good evidence that this would have worked.
But it they could not or would not do it.

Is a community notes system going to do better? There are important questions of timeliness and reach. However, research shows community notes both
1) produces objectively good results, and
2) maintains trust in the people being fact checked
academic.oup.com/pnasnexus/arti…

Again, it's not good that Meta has stopped working with pro fact checkers. Although they only ever checked a tiny minority of content, having dedicated humans watching what's going viral is a very valuable service.

But this experiment failed. It could not generate trust.

/fin