AI Will Reshape Public Opinion (www.conspicuouscognition.com)

Recent discussions within the AI/ML community highlight the potential of large language models (LLMs) to reshape public opinion by reversing the democratizing effects that social media has had on discourse. While social media has allowed ordinary voices to bypass traditional gatekeepers, resulting in the amplification of populist and sensationalist messages, upcoming technologies like OpenAI’s GPT-5 and others are being positioned as expert-level tools that could enhance public access to reliable, evidence-based information. This shift is significant as it could lead to increased trust in expert opinion, potentially countering the misinformation and division often seen in today's media landscape. Key technical details reveal that LLMs, through their extensive training and data integration, can provide well-rounded, factual responses to various topics when prompted, outperforming typical social media narratives. This capability allows LLMs to serve as "epistemic intermediaries," similar to the vision proposed by early 20th-century journalist Walter Lippmann, who advocated for structured avenues of expert knowledge. As LLMs become more integrated into public discourse, they may not only provide answers aligned with expert consensus but also reshape how individuals perceive and organize political and societal issues, moving society towards a more informed public while posing new questions regarding control and influence in this evolving information landscape.