Facebook, others made demonstrable progress in combating misinformation during the midterms but the problem is not going away due to engagement-based algorithms
but only within the bounds of a core product and business model that inherently lend themselves to manipulation. http://slate.com/... http://twitter.com/... Will Oremus / @willoremus : A big question going into the 2018 midterms was whether platforms like Facebook and Twitter could fix their issues with rampant misinformation and foreign meddling. And the answer turned out to be... not really! Those problems are here to stay. https://slate.com/... pic.twitter.com/koxCYNaSzR Erin Gallagher / @3r1ng : “Our political conversations are now happening on an infrastructure built for viral advertising, on platforms that are purpose-built to generate engagement and amplify sensational content” - @noUpside via @cwarzel http://www.buzzfeednews.com/ ... Will Oremus / @willoremus : From now on, the question is not *whether* social media will be gamed to manipulate elections, but *how badly.* And the answer in 2018 was, well... it could have been worse! http://slate.com/... Matthew Yglesias / @mattyglesias : But this just means they're not trying! http://twitter.com/... Kevin Roose / @kevinroose : Yup. As long as the model is limitless sharing + engagement-ranked feeds + self-service ads, all the war rooms in the world won't do much. http://twitter.com/... Will Oremus / @willoremus : Now the good news: Thanks to pressure from the public, Congress, and the media, the big platforms made real progress on almost every front. http://slate.com/... http://twitter.com/... Matthew Yglesias / @mattyglesias : I take a tough line on this because it appears to be important to these companies' talent retention strategies to have the reputation that they are trying to solve the misinformation problem when they are in fact committed to a core business model that makes it impossible. Charlie Warzel / @cwarzel : New from me: One thing the information war surrounding the lead up to midterms revealed is that the platforms have already rewired our political discourse and we all have brain worms now http://www.buzzfeednews.com/ ... http://twitter.com/... @slate : Twitter and Facebook were still a garbage fire in the midterms. But at least now there are firefighters. http://slate.com/... http://twitter.com/...
Context & Ripple Effects
Going into Election Day, Jonathan Albright's analysis of 250K posts and 5K political ads had already documented how much politically motivated misinformation was circulating on Facebook ahead of the midterms. Will Oremus's verdict after the vote splits the difference: the platforms' cleanup efforts were real, but they operated inside a product architecture — engagement-ranked feeds and self-service ads — that rewards exactly the behavior moderators are trying to suppress.
The piece sets up a pattern the corpus keeps confirming: two years later, platforms had again made progress since 2016 on the most blatant abuses, yet reviewers still judged the problem endemic to the product rather than a fixable bug.
First-order effects
- Facebook and Twitter's trust-and-safety teams get credit for measurable midterms gains, but the same companies must now defend an ad business whose self-service targeting is the manipulation vector itself.
- Campaigns and political advertisers learn that viral distribution remains purchasable, keeping cheap influence operations viable despite post-2016 safeguards.
Second-order effects
- Rivals face the same bind when they act: YouTube's later crackdown on fringe-channel recommendations simply shifted its algorithm's favor toward mainstream partisan outlets like Fox News, showing suppression upstream relocates rather than removes the problem (per the NYT study).
- Platforms respond by staging visible crisis infrastructure — the election 'war rooms' later dismissed as PR theater — converting moderation work into a communications exercise aimed at press and regulators.
Third-order effects
- If every intervention stays bounded by engagement economics, the endgame is external: Congress and regulators become the only actors positioned to force changes to ranking and ad-targeting systems the platforms won't make themselves.
- Warning-label and labeling regimes harden into standard practice while doing little to the underlying supply of misinformation — [[a:959842|the Washington Post called the 2020 labels a public-relations win that came too late as a product fix]] — entrenching a cycle where optics outpace architecture.
The trend: Platform moderation improves at the margins each election cycle while engagement-based distribution keeps the core vulnerability intact, steadily shifting the debate toward regulatory intervention.