An optimistic – or maybe fatalistic? – take on how the increasing presence of AI in the information ecosystem will shape public opinion. In a nutshell: When we look at large language models as flawed information transmitters, we’re asking the wrong question. The relevant question is «How do they compare against alternative sources of information?» And the author argues: «LLM’s responses are generally much more accurate, evidence-based, and in line with expert consensus than what you get from most social media posts.» It’s an interesting perspective – that doesn’t take into account that much of the dis- and misinformation circulating on social media is now supercharged with large language models.

From Weekly Filet #577, in July 2026.