Why Widespread LLM Use Could Become a Collective Disaster

Summary

The essay argues that large language models may become a new collective-action failure: tools that save individuals and companies substantial time could make life worse for everyone when adopted broadly. The author says the time saved by writing, research, coding, and idea development is unlikely to become leisure because AI-assisted people will compete more intensely for scarce attention, making effective AI use a baseline requirement rather than a lasting advantage. The article also identifies digital trust as a public good being eroded. Long documents, workplace requests, and carefully reasoned submissions may receive less attention when anyone can cheaply generate large volumes of polished text. As an example, the author describes a tool that could analyze UK planning applications and automatically produce thousands of words of objections, potentially overwhelming council officials and reducing attention for legitimate concerns. The essay connects this pattern to the weakening of academic publishing, where individual optimization can damage the wider system. It does not propose a definitive remedy, noting that spam detection tends to lag behind content creation. The author suggests that people may eventually rely mainly on trusted sources, and regrets the possible loss of an internet where strangers could engage meaningfully.