This is about to be some real intense Facebook distrust, which is of course more warranted after all of the prior events that occurred during the last election.
It appears to me that Facebook has formed an evolved strategy to manipulate elections. There I was, flicking through Facebook on my phone and a Bernie Sanders advertisement appears. It is focused on climate change and I noticed something very peculiar going on in the comments section: swarms of vegans posting in response, criticizing Bernie Sanders for this advertisement. At first glance I was under the impression that veganism became popular enough that a lot of the commenters were pro-vegan. Upon further investigation I found this new feature that could aide in explaining this mystery: Most Relevant, a new way to organize comments on mainstream posts with massive comment sections. This along with Newest and a few other organizational algorithms have been recently added to Facebook.
This really started to worry me, as it’s clear this is the perfect way to socially engineer the next election by creating demographic-specific lists of filtered opinions that bias people using their conformity and tribalism in favor of Facebook’s political goals. I was primarily seeing vegans reject Bernie Sanders and concerningly, I am vegan. The comments were nitpicking Bernie for not being vegan, many saying they can’t vote for him because of this. This seemed very strange as I suspect most vegans I know would not be so petty as to inadvertently support Trump by dividing liberal candidates based on whether or not they are vegan.
Next I searched Bernie Sanders page in order to investigate further. One post on LGBTQ stuck out. The general trend in this comment section was full support for Bernie Sanders. My sexual orientation is listed as straight on Facebook and generally speaking it is straight individuals who are most likely to be homophobic (duh) and so the most reliable tribal bias to show me, a non-LGBTQ tribe member, of the LGBTQ tribe’s opinion is a pro-Bernie bias. If I am straight, one could assume I have little investment in the opinions of LGBTQ tribe members and if I do have a strong relevant bias in response to LGBTQ member’s opinions as a straight individual it would likely be anti-LGBTQ. So by showing me pro-Bernie commentary coming from the LGBTQ community it can either have little to no effect due to personal/selfish irrelevance or bias me against Bernie due to homophobia. To be clear, I am not homophobic and may even be slightly bisexual or at least bi-curious (I’ve heard people get teased for this label) but I have myself listed as straight on Facebook either way.
In another post on Bernie’s page, there was mention of 20-40 age demographic in support of Bernie’s politics. Strangely, the most relevant comments in this case were showing 50+ agers. It gave off the vibe that old people are going to exploit my taxes because they are getting increasingly sick due to aging. Somehow I couldn’t even find commenters my age, that is until I switched to the Newest algorithm for comment listing and scrolled way down to find a 25 year old who says they won’t be fooled into this trap. The very little comments I found in my demographic were biasing against Bernie, continuing this pattern we’ve observed thus far.
In another post about increasing teachers pay, there were many commenters noting that they couldn’t believe the negativity in the comments and yet I struggled around looking for negative comments to no avail under the Most Relevant order of comments. It could be that these people were lumped into the demographic that is supporting Bernie, as they clearly are, while the negative ones are filtered out from my demographic’s perspective. The people who are observing a plethora of negative comments might be on the other side of the demographic fence, where they are being fed anti-Bernie sentiment from their own tribal affiliation while I’m being streamed an ‘enemy’,or outsider tribal sentiment in hopes that I will also become anti-Bernie.
It’s worth noting that this sort of manipulation would be incredibly easy to conceal if the algorithms weren’t fixe, in other words, a dynamic algorithm. One possible method of increasing noise and undetectability would be to have increasingly detailed sub-demographics, for example vegan liberal or vegan conservative, or worse if Facebook uses ‘Bernie supporters’ as an actual tag in the Most Relevant algorithm. Using this sort of label, combined with others that resonate more heavily to one’s personal identity than candidate support does could have the power to sculpt social norms and lead tribal conformity in favor of Facebook’s biases. As mentioned, a dynamic algorithm could camouflage this social engineering by alternating the demographic tags that Most Relevant comment listing filters for as to eliminate any trackability. Some days the algorithms could be designed for social control while other days they could be innocent programs. It’s also possible that Facebook could identify and label vulnerable and suggestible individuals so that they could narrow down the amount of people exposed and thus reduce the risks of being caught.
Another valuable target of influence would be individuals who many others conform to. If Facebook can identify a sort of leader trait demographic then these individuals may be of high value because altering their perceptions would impact large swaths of people and even further reduce the chances of Facebook getting caught up in legal scandals. The most obvious would be to manipulate celebrities but it may be valuable to find less popular influencers as well. Likely this would just boil down to individuals whose posts have large amounts of engagement. Also they could detect where social contagion patterns originate. All of this suggests to me that it would be easy for Facebook to engineer political culture while masking themselves in a sea of social noise.
Another clever way to alter culture would be to boost the reach of posts that help Facebook’s political goals. Even boosting opposing opinions could be valuable because they could create the image that the boosted tribe is offensive. Facebook could detect which liberals post offensive or bigoted social justice content, which may have happened already in the past. By giving SJW culture cooties, Facebook could seem to serve as a platform for liberal ideas while also boosting the most ugly side of their statements, actually destroying their identity with stigma and preventing the group from expanding as much due to people’s fears of being persecuted by association. Antifeminism and anti-SJW are both newly emergent reactionary cultures, and obviously Trump is a most radical symbol of all of this.
Ultimately these algorithms may serve to disintegrate consensus among certain tribes as to diminish their ability to have democratic influence on society. This reminds me of the way YouTube can cause people to spiral out into hyper-specific content. YouTube is like a pick your path game that engineers your ideologies based on suggested content that continues to escalate down the pipeline until you become radicalized at the hands of an algorithm. With Facebook it seems we are seeing the merge of this algorithmic identity tunneling and political social engineering.
It’s totally possible that I am experiencing some apophenia but regardless even if we aren’t being intentionally programmed, at least this reveals the potential power that exists in this new feature, which is fascinating on its own. This also brings to light the possible novel societal dynamics that may occur naturally due to this ‘most relevant’ comment organization. Have you noticed patterns like this? Pay attention to your demographic affiliation and the way that the Most Relevant algorithm feeds you memes and ideas. After looking around Facebook a bit please report your experience in the comments below.
Special thanks to the two patrons, Abhishaike Mahajan and Charles Wright! Abhi is also the artist who created the cover image for this post. Please support him on instagram, he is an amazing artist!
If you liked this, follow me on
If you’d like to support these projects check out this page.