A counter-point to this is that studies show polarization has also fallen in some countries over the past years - including ones where social media (Facebook or otherwise) is popular. Studies also show some of the most polarized segments in the US to be the older population, which uses social media less. We definitely have work to do, but this suggests there are many factors at play.
Yep we actually do this! (invite journalists to decision-making meetings). One of our regular meetings is about the content policies, we publish the minutes here - https://about.fb.com/news/2018/11/content-standards-forum-mi... - and have also hosted journalists and outside academics from time to time.
Might be a long list. We've got >35,000 people working on safety and security. I expected some challenges recruiting people with all the bad press we got over the last few years, but instead I've seen that talented people eager for a really hard challenge are even more excited to join and wok on this area.
Fair enough, but I also get to see first-hand how decisions are made and how rigorous debates take place, so I have more faith in the process. We've got lots to do to improve transparency of how this stuff happens, because I know people care about it. One way we started a while ago is publishing minutes to one of our meetings where decisions on content policies get made. https://about.fb.com/news/2018/11/content-standards-forum-mi... -- lots more to do!
If we're talking metrics, the 64% stat cited in the article turns out not to be a good way to measure impact of recommendations on an extremist group. We internally think about things like prevalence of bad recommendations. More on prevalence here - https://about.fb.com/news/2019/05/measuring-prevalence/. I don't have a metric I can share on recommendations specifically but you can see the areas we've shared it for so far here: https://transparency.facebook.com/community-standards-enforc...
Hey, Facebook VP here (I work on this). We’ve made some meaningful changes to address this since 2016. We’ve strengthened our enforcement in groups and have been actively working on our recommendation tools as well, for example removing groups with extremist content that violates our policies, from recommendations.
Hey, Facebook VP of Integrity here (I work on this stuff).
This WSJ story cites old research and falsely suggests we aren’t invested in fighting polarization. The reality is we didn’t adopt some of the product suggestions cited because we pursued alternatives we believed are more effective. What’s undeniable is we’ve made significant changes to the way FB works to improve the integrity of our products, such as fundamentally changing News Feed ranking to favor content from friends and family over public content (even if this meant people would use our products less). We reduce distribution of posts that use divisive and polarizing tactics like clickbait, engagement bait, and we’ve become more restrictive when it comes to the types of Groups we recommend to people.
We come to these decisions through rigorous debate where we look at all angles of how our decisions will affect people in different parts of the world - from those with millions of followers to regular people who might not otherwise have a place to be heard. There’s a baseline expectation of the amount of rigor and diligence we apply to new products and it should be expected that we’d regularly evaluate to ensure that our products are as effective as they can be.
We get criticism from all sides of any decision and it motivates us to look at research, our own and external, analyze and pressure test our principles about where we do and don't draw lines on speech. We continue to do and fund research on misinformation and polarization to better understand the impact of our products; in February we announced an additional $2M in funding for independent research on this topic (e.g. https://research.fb.com/blog/2020/02/facebook-misinformation...).
Criticism and scrutiny are always welcome, but using cherry-picked examples to try and negatively portray our intentions is unfortunate.