Of the 20 most-engaged Facebook posts containing the word "election" over the past week, all of them (literally 100%!) are from Trump and have labels indicating that they're false or misleading.

What are the labels doing, exactly?

Meanwhile, YouTube videos containing false voter fraud allegations have been viewed *rubs eyeballs* 138 million times according to one estimate. https://www.nytimes.com/2020/11/18/technology/election-misinformation-often-evaded-youtubes-efforts-to-stop-it.html

Voter fraud conspiracy theories pushed by a sitting president are a bigger problem than social media companies alone can solve, but man, ranking information based on how interesting it is has consequences. https://twitter.com/_cingraham/status/1329515194121289733

The problem with tech companies "fighting" misinformation is that false things are generally way more interesting than true things. If your system is built around an engagement-ranked feed, you can label and fact-check all you want and it won't move the needle much.

Sidenote: this argument (that engagement rises as content gets worse, barring manual intervention) was made a few years ago by a promising young data scientist, "M. Zuckerberg." https://www.facebook.com/notes/mark-zuckerberg/a-blueprint-for-content-governance-and-enforcement/10156443129621634/