A new analysis of thousands of Donald Trump’s social media posts has revealed a pattern that should give pause to anyone concerned about the health of democratic discourse. Researchers at Stanford’s Digital Democracy Lab fed a dataset of over 20,000 tweets from Trump’s official account (2015-2020) into a natural language processing model to track shifts in tone and content. The results show a clear and unsettling trend: a steady increase in authoritarian language, personal attacks, and reinforcement of misinformation, with a corresponding decline in substantive policy discussion.
The study, published in the Journal of Computational Social Science, identifies a ‘normalisation effect’: as the former president’s rhetoric escalated, so did the language of his followers and even mainstream media outlets, effectively shifting the Overton window of acceptable political speech. Lead researcher Dr Elena Vasquez warns that this is not just about one individual. ‘We’re seeing a feedback loop between elite discourse and public sentiment.
The algorithms that amplify this content are optimised for outrage, not truth.’ The findings highlight the urgent need for digital sovereignty: a framework where citizens control their data and platforms are held accountable for algorithmic amplification of harmful content. For democracy to survive the quantum age, we must recalibrate the user experience of public debate.
