This research addresses the political role of social media influencers. As trust in the mainstream media continues to decline, influencers - ranging from lifestyle content creators to social mobilisers and topical experts - have become a source of information for many. Existing research has yet to study the ramifications for society on a large scale, failed to isolate the role of influencers from that of users and platform algorithms in driving the popularity of (or suppressing) influencer content, and importantly, failed to take a cross-platform approach. Combining computational methods, panel surveys, experiments, and social media takeout data, I advance a novel theoretical model of influencer politics. I build an integrative typology and comprehensive map of social media influencers, their content, and audiences, and assess whether some populations are more at risk of consuming (harmful) influencer content. Further, using several data sources, I isolate the editorial role of social media platforms and their algorithmic systems in content curation, examining whether these algorithms amplify voices and silence others. Finally, through an experimental study, I investigate the conditions under which influencers shape political attitudes and online behaviour. My study encompasses five major platforms, a necessity considering cross-platform variations in user-bases and algorithmic systems. My research relates to three SEDarc themes: transformative technologies for society; healthy, thriving communities; and secure, effective, and trusted institutions, by answering whether AI-tools have a harm-bias and whether influencers can foster political engagement, integrate diverse groups into democratic discourse, and contribute to improving trust in democratic institutions and processes.