1 Matching Annotations
  1. Nov 2023
    1. Firstly, we must continually question the underlying assumptions, potential pitfalls, risks, and possible unintended adverse effects of introducing AI into democratic processes. Not the least by always checking and refining LLM outputs with real people, or we risk falling into the fallacies and risks of democracy in silica.

      Where here is the role of AI? And does it matter as much in each of its roles? - moderation of conversation - synthesising new statements (this one particularly?) - transcripts - summarising opinions - determine statistically more supported statements - the import of minority statements? (e.g. all may have an opinion, maybe not all opinions matter the same way in a case (democratically built bridges may fall, in comparison one built by engineering teams) --> this points to curating the issues to discuss. And ensuring all voices are indeed weighed, not just outvoted, such that groups aren't marginalised.