The Role of Network Diffusion and Information Cascades in the Spread of Misinformation
Both fearing and despising yellow journalism, Adolph Simon Ochs became one of the most prominent and esteemed newspaper publishers of all time for his unwavering commitment to journalistic neutrality and high standards of ethics, which can be summed up in his ever-enduring slogan and approach to newspaper publication: that he only publishes “All the news that’s fit to print.” His push towards ethics and journalistic neutrality inspired many generations of publishers to come, and played a pivotal role in preventing much of the wildly-divisive partisanship that marks current political attitudes. However, the rise of social media has allowed for actors to bypass normal channels of information dissemination, because social media cuts out the middleman—preventing any exercise of editorial discretion—and allows actors of all backgrounds and motivations to directly share unregulated information, at no cost to the actors. With limited barriers to entry and a combined audience of billions, social media has become, by far, the best conduit for misinformation.
The problem with social media information dissemination is two-fold: not only are social media companies providing sources of misinformation with a platform of billions of active users, but its reliance and dependency upon deep-learning neural networks—which provides its users with tailored content—also creates widespread network diffusion and further divides people along ideological lines.
When information is shared on Twitter by way of tweets, it makes many impressions on many different user’s pages: those who follow the original poster, followers of a person who retweets/replies to the tweet, as well as accounts Twitter’s artificial intelligence algorithms deem will be interested in the information. Since these interactions and relationships are mostly the result of personal and ideological similarities, it creates a somewhat-tightly connected network that is capable of demonstrating information cascade and network diffusion effects. That is why social media networks, and especially ones where AI tailors its content to users, are known as echo chambers.
Since users are all ostensibly connected along ideological lines, when one user in the network picks up and shares misinformation, other users in the network are also likely to share and uptake this same misinformation, because the strong ties among followers in addition to misinformation that fulfills an ideologically-satisfying narrative makes the misinformation more salient to the users in the network. When the information is salient, more users are likely to uptake and spread it. As more users in the network uptake and share the misinformation, it creates a cascading effect. If a user in a network sees only two narratives—truth and misinformation—and the misinformation is legitimized by close-tie friendship as well as provides the user more utility by way of an ideologically-appealing narrative, it is quite likely that the user, too, will adopt and share the misinformation. As more users in the network adopt the beliefs, it only further legitimizes the misinformation. Since misinformation is capable of better resonating with users, it creates a perverse incentive to continue creating and sharing misinformation.
The information cascade and network diffusion effects associated with misinformation dissemination on social media have been particularly effective in the context of domestic politics and partisanship; the discord sown by misinformation can be seen just by glancing at the President of the United States, in addition to practically any other elected official. The uptake and spread of misinformation is particularly concerning because it creates an unbridgeable domestic divide, destabilizes the United States and its populous, as well as opens the floodgates to manipulation by foreign actors.