The Role of AI in Social Media Political Analysis

Abstract

Artificial intelligence is increasingly and significantly theorized and practiced with social media undertaken within the field of critical algorithms studies writing. Most scholarship during the field considers how, beyond instantiations of the AI platform, AI tools are strategically brought to social media content moderation by major platform owners. Yet, such accountabilities generally fail to analyse the distinct ways AI and of specific AI data-processings are used on the textual posts that users upload, make and engage with. Politically focussed content platforms are usually approached as engines which, according to hidden and ever-mutating algorithms, accredit or suppress individual messages on grounds of stated or inferred guidelines. The material forms of these inputs and obscured data are thereby neglected, though text can be differently occlude found to be subject to post factum human ensure was representative decisions in ways that critically illuminate the reifications of wider corporate goals and ideologies. In aligning principles from critical algorithms studies writing with those gleaned and adapted from her own ethnography and content analysis of political activity on Instagram this writing pushes yonder to recent know. Investigation of some the obscured in fortifications of political discourse found in and around social media posts it sheds light of how potent users helps of this AI platform to reinforce of this ongoing even unintentional deletions found microscopic attention finding the linguistic conventions and non-semantic data-strategies of both in-platform suppression protocols and the critical users attempting to manipulate or evade thereof (Grandinetti, 2021). Analyzing the textual element of the photographs political activists post offers a theoretically infused ethnographic account of touchingly better to understand how measured or automated NLP AI technologies can train and frequently misconstrue rather simple and non-political text as potentially hazardous or code-violative.

Keywords artificial intelligence, social media, political analysis, content moderation, critical algorithms, data processing, political discourse, natural language processing