diff --git a/contents/english/5-4-augmented-deliberation.md b/contents/english/5-4-augmented-deliberation.md index 20ad1e59..c08bc5fa 100644 --- a/contents/english/5-4-augmented-deliberation.md +++ b/contents/english/5-4-augmented-deliberation.md @@ -87,7 +87,17 @@ Other more experimental efforts, closely aligned with the techniques discussed i Some of these more ambitious experiments begin to point towards a future, especially harnessing language capabilities of GFMs to go much further towards addressing the “broad listening” problem, empowering deliberation of a quality and scale that has henceforth been hard to imagine. The internet enables collaboration at an extreme scale by reducing the possible space of collaborative actions, such as to buy/sell market transactions, and by utilizing a similar reduction in information transmission, i.e. to five star rating systems. An effective increase in our ability to transmit and digest information can result in a corresponding increase in our ability to deliberate on difficult and nuanced social issues. -One of the most obvious directions that is a subject of active development is how systems like Polis and Community Notes could be extended with modern graph theory and GFMs. The "[Talk to the City](https://ai.objectives.institute/talk-to-the-city)" project of the [AI Objectives Institute](https://ai.objectives.institute/), for example, illustrates how GFMs can be used to replace a list of statements characterizing a group's views with an interactive agent one can talk to and get a sense of the perspective. Soon, it should be possible to go further, with GFMs allowing participants to move beyond limited short statements and simple up-and-down votes. Instead, they will be able to fully express themselves in reaction to the conversation. Meanwhile, the models will condense this conversation, making it legible to others who can then participate. Models could also help look for areas of rough consensus not simply based on common votes but on a natural language understanding of and response to expressed positions. A recent large-scale study highlights the positive impact of such tools in enhancing online democratic discussions. In this experiment, a GFM was used to provide real-time, evidence-based suggestions aimed at refining the quality of political discourse to each participant in the conversation [^LLMDemocracy]. The results indicated a noticeable improvement in the overall quality of conversations, fostering a more democratic, reciprocal exchange of ideas. +One of the most obvious directions that is a subject of active development is how systems like Polis and Community Notes could be extended with modern graph theory and GFMs. The "[Talk to the City](https://ai.objectives.institute/talk-to-the-city)" project of the [AI Objectives Institute](https://ai.objectives.institute/), for example, illustrates how GFMs can be used to replace a list of statements characterizing a group's views with an interactive agent one can talk to and get a sense of the perspective. Soon, it should be possible to go further, with GFMs allowing participants to move beyond limited short statements and simple up-and-down votes. Instead, they will be able to fully express themselves in reaction to the conversation. Meanwhile, the models will condense this conversation, making it legible to others who can then participate. Models could also help look for areas of rough consensus not simply based on common votes but on a natural language understanding of and response to expressed positions. + +Such cutting-edge approaches are beginning to appear not only in policy deliberations and community dialogues but also in electoral processes. In the 2024 Tokyo gubernatorial election, candidate Takahiro Anno leveraged Google Slides and GitHub to publish his manifesto[^AnnoManifest], invited voter feedback via X (formerly Twitter), Google Forms, AI-handled incoming call and GitHub, visualized perspectives using Talk to the City[^AnnoTTTCReports], and engaged in policy refinements through GitHub discussions[^AnnoGitHub]. Moreover, by deploying an AI virtual avatar(AI-Anno[^AIAnno]) on YouTube that responded to questions around the clock, he demonstrated a interactive form of “broad listening.” Despite being a relatively unknown candidate, Anno garnered over 150,000 votes, suggesting that the technological potential exemplified by Talk to the City and the participatory, deliberative qualities of platforms like vTaiwan and Polis can extend meaningfully into the electoral sphere[^Anno2024]. + +[^AnnoManifestSlide]: [Manifest on Google Slide](https://docs.google.com/presentation/d/1kE_W3NpvIODglaN1OrKKxmq-rNMu69Vnh7M9hhSpFwU/edit#slide=id.g274ffc7e6e4_1750_3430) and [GitHub](https://github.com/takahiroanno2024/election2024) +[^AnnoTTTCReports]: [Talk to the City Reports](https://takahiroanno2024.github.io/tokyoai-analysis/) +[^AnnoGitHub]: From June 21 to July 6, over 15 days, 232 issues were raised, 104 proposals for change were submitted, and 85 were adopted. +[^AIAnno]: During a 16-day period, the AI-Anno answered approximately 7,400 questions. This represents 77% of the submitted inquiries, significantly surpassing the capacity of a single human respondent in a lecture format with one speaker and multiple viewers. +[^Anno2024]: 150,000 votes accounted for 2.3%, marking the highest number of votes in history for a candidate in their 30s in the 22 past Tokyo gubernatorial elections. Anno finished fifth rank. Following this gain in recognition, a special election program on October 15 for the House of Representatives used Talk to the City to confront incumbent politicians with citizens' concerns. Starting November 22, he has used Talk to the City for a project to gather public input for Tokyo's long-term plan toward 2050. As of December 13, over 10,000 opinions have been collected, far exceeding the typical response to public comment calls. + +A recent large-scale study highlights the positive impact of such tools in enhancing online democratic discussions. In this experiment, a GFM was used to provide real-time, evidence-based suggestions aimed at refining the quality of political discourse to each participant in the conversation [^LLMDemocracy]. The results indicated a noticeable improvement in the overall quality of conversations, fostering a more democratic, reciprocal exchange of ideas. While most discussion of bridging systems focuses on building consensus, another powerful role is to support the regeneration of diversity and productive conflict. On the one hand, they help identify different opinion groups in ways that are not a deterministic function of historical assumptions or identities, potentially allowing these groups to find each other and organize around their perspective. On the other hand, by surfacing as representing consensus positions that have diverse support, they also create diverse opposition that can coalesce into a new conflict that does not reinforce existing divisions, potentially allowing organization around that perspective. In short, collective response systems can play just as important a role in mapping and evolving conflict dynamically as helping to navigate it productively.