Artificial Intelligence Makes Fake News More Credible

Source Domain: www.miragenews.com

Here’s an unordered list summarizing the significant points from the article on the linguistic features of fake news and developing fact-checking tools:

Language Project “Fakespeak”: Linguist Silje Susanne Alvestad led the ‘Fakespeak – the language of fake news’ project, studying how the language characteristics of fake news can be identified in English, Russian, and Norwegian.
Linguistic Peculiarities of Fake News: The research discovered linguistic markers of fake news, such as:
- Predominant use of present tense indicative of falsehood
- Informal, conversational style
- Higher usage of emphatic expressions
- A more categorical and certain tone
Impact of Writer’s Motivation: Fake news varies linguistically depending on the writer’s motivation:
- Monetary motivations correspond to fewer metaphors
- Ideological motivations correlate with more metaphors, especially from sport and war
Challenge for Fact-Checking Tools: The linguistic features of fake news vary widely and are influenced by individual languages and cultural contexts, making it challenging to create a universal fact-checking tool based purely on linguistic characteristics.
NxtGenFake Project: Research on the linguistic features of AI-generated disinformation complements the ongoing project “NxtGenFake”, which is focused on AI-generated disinformation.
Characteristics of AI-Disinformation:
- Less variation in persuasive techniques
- Frequently employs an “Appeal to Authority”
- Generally concludes with “Appeal to Values”
Public Perception of AI-Generated Disinformation: In tests, the AI-generated disinformation was rated as more credible, more informative, and preferred over human-written disinformation, indicating a potentially greater challenge in detection.