Meta wants to get politics out of its platforms. Is this good or bad for democracy?
For Meta, politics is bad for business. And for two reasons. First, politically charged content and the toxicity surrounding it do not create an attractive space for advertisers. In late 2023, IBM pulled its ads out of Twitter after discovering its brand appeared next to pro-Nazi content. Second, there is only one existential threat for Meta – and it is not competitors but politics. The US government can break the Internet platform apart or fatally hurt its business model. Democrats have openly contemplated breaking up plans for Meta, while Republicans have attacked Meta content moderation policies and their supposed left-wing bias.
Meta has produced a number of strategic decisions in early 2024 that aim at steering the company clear of politics.
First, with a decision that has no impact on the experience of its users but with profound consequences on how the company’s platforms are perceived in the public eye, Meta has announced that it will finally shut down Crowdtangle, a search tool used by researchers and academics to get a quantitative, bird-eye view on Facebook and Instagram and their users’ conversations. An experiment in transparency co-founded by Brandon Silverman (who left Meta in 2021), Crowdtangle was a continuous source of embarrassment for Meta and a PR nightmare. Second, Meta has recently confirmed that on Threads and Instagram, it will not “proactively recommend political content from accounts [that users] don’t follow,” while on Facebook, it will maintain a policy of not recommending political content based “on engagement signals”, using instead an array of “personalized signals”. The recent decisions to downgrade news on their platforms should also be placed in a broader content realignment strategy.
Meta executives seem to think that the less political attention the company gets from its users and policymakers, the better.
A realignment away from politics toward other interests might make sense for a for-profit organisation. Still, since social media are a critical venue for the public to be exposed to politics and engage with politics (in 2023, 30% report social media to be the main way of getting news), we must ask: are Meta’s decisions good or bad for the quality and functioning of our democracies?
Social media companies did not design their platforms around content moderation. Instead, they have been progressively forced into developing content moderation policies and interventions by public outrage. Yet they have always struggled at the task. In our own research on the online performance of the far-right community during the pandemic, with Amelia Johns, Emily Booth and Marian-Andrei Rizoiu, we found that Twitter’s aggressive COVID-19 policies on harmful content did not prevent Australian far-right accounts from reaching at the onset of the pandemic significantly more users with their message than the 2019-20 bushfire seasons. Studying anti-vax groups on Facebook before and after the COVID-19 outbreak, we found that if Meta managed to reduce the reach of anti-vax content from selected groups throughout its platform, it could not prevent some of the anti-vax pages from expanding a few times over their audiences during the pandemic.
The failure of social media policies to rein in the diffusion of harmful content (for both users and brands) during the information crisis following the COVID-19 outbreak, but also the sustained attention of the political establishment towards the company since the 2016 US presidential election (Mark Zuckerberg appeared before the US Congress in 2018 and 2021), seems to have suggested to Meta that containment instead of suppression was the only financially as well as politically effective solution to content moderation. Under this approach, harmful and politically charged content can be tolerated but not surfaced to average users by segregating the creators of political content – including news creators – to the rest of the user base.
The effect of social media on democracy is complex and certainly not uniquely positive. Engagement-based recommender systems distort the bandwidth of political communication by promoting content that triggers strong emotional responses. However, social media are sometimes the only available resource for political participation for those who need access to other resources because of their socio-economic status. Social media have been found to improve political engagement among those who would not be otherwise exposed to politics, bringing political content to those who are disengaged and not actively seeking it.
The accidental exposure to political opinions we disagree with is never a pleasant experience. But so it is to find out our candidate has lost an election. Democracies allow diverse values to coexist peacefully. Critical for this coexistence is the institutionalisation of continuous processes for coordinating interests held by widely diverse citizens. Votes such as elections and referenda are, in fact, an important moment of public reckoning - when this diversity is exposed and counted. But voting is limited in how much information can communicate about the electorate. Complex societies need communication channels that can convey complex sets of interests. Without horizontal communication, democratic societies lose their capacity to coordinate, accommodate, and change.
When Meta’s social media platforms, arguably the most important piece of infrastructure for horizontal communication, demote political content because it is bad for business, either democratic society opens up alternative communication channels, or they necessarily become less democratic.