As a rule, socialbots are designed to pass the Turing test: They're sophisticated enough to fool other users and be taken for a human. To do so, socialbots use artificial intelligence (AI), text mining and data analysis software. Some have access to databases of general knowledge information and current events to allow them to recognize references and craft more convincing messages.
Socialbots are usually programmed to interact as a human does, following relatively normal sleep/wake cycles and sending messages on a somewhat randomized schedule rather than at regular intervals. Twitter is the most common platform for socialbots (see: Twitterbot) because the microblogging format is well-suited to the software's capabilities.
Most socialbots are created for a specific purpose, such as marketing, political campaigning or public relations. Researchers from the Federal University of Ouro Preto in Brazil created a socialbot they called Carina Santos for a journalistic Twitter account. By the time the researchers revealed that the popular account was a socialbot, the faux journalist outranked Oprah Winfrey on Twitalyzer, an influence-rating website.
The capacity of socialbots to wield influence could enable them to sway voters, mount political attacks or overwhelm dissent, among many other possibilities. Socialbots can also pose a security risk. In 2011, for example, a socialbot network stole gigabytes of user data from Facebook.
According to some estimates, up to 65 percent of the average Twitter account's followers are actually bots of one type or another.