Technologies of truth on the web
A database-trained artificial intelligence concludes that messages are false, without resorting to comparison with true judgments and logical procedures, because of the economy of cognitive effort.
Abstract
The spread of misinformation on the Internet today is a serious problem due to the vast impact on society of network information. The most obvious technology aimed at combating false information is fact-checking, which allows to identify the presence of facts in the message and compare them with the base of true information. Such technologies are applicable for tracking information distortions, but do not allow evaluating a random message. An alternative approach is to identify false messages based on indirect signs: its linguistic and paralinguistic features, as well as on the communicative history (author, creation and distribution) and other features. Database-trained artificial intelligence concludes that messages are false, without resorting to comparison with true judgments and logical procedures. The ability of a person to independently evaluate the truth of messages is limited by the economy of cognitive effort, and people are even able to generate their own memories that confirm false messages.