[Barrin92]

Yes, we have that research from Facebook itself, which they promptly shelved.

https://www.theverge.com/2020/5/26/21270659/facebook-divisio...

"“Our algorithms exploit the human brain’s attraction to divisiveness,” one slide from the presentation read. The group found that if this core element of its recommendation engine were left unchecked, it would continue to serve Facebook users “more and more divisive content in an effort to gain user attention & increase time on the platform.” A separate internal report, crafted in 2016, said 64 percent of people who joined an extremist group on Facebook only did so because the company’s algorithm recommended it to them, the WSJ reports."

[im3w1l]

The Verge cites WSJ. WSJ mentions "a presentation" but doesn't include it. What little they cite is not very substantive. "more and more divisive content" this is vague and omnious, and not even very likely. These algorithms will reach an equilibrium.

WSJ further mentions that "some proposed changes would have disproportionately affected conservative users and publishers"

To me this is a sign that the presentation was made by a partisan inside of Facebook. All in all, I don't think this is very confidence inspiring.

We would really need some third party to do a better job of investigating this.

[Barrin92]

You gotta be kidding me. You know as well as I do that we're never going to get comprehensive third party access to Facebook data.

When even internal research suggests that their platform drives division your conclusion is they're overstating their case, people who are on facebooks payroll?

That's like Enron publishing internal research about corruption and someone going "well I guess these guys are partisan"

[im3w1l]

I've worked in ranking myself and seen how the debates go from the inside. I would not take a conclusion like this at face value, no. If they had included some data, maybe.

What you would likely have to do to determine this as an outsider is sample a number of people, ask them for opinions on stuff, look at their feeds. See if certain opinions are overrepresented. Look at correlation between opinions they have and opinions they see while you are at it.