moda 官網 CIP 資訊

1. 民主司交付期限：11/28
2. 策略司：政大團隊已於12/29完成中文內容並同步進行英譯，英譯內容將配合中文內容滾動調整。
3. 網址：moda.gov.tw/major-policies/alignment-assemblies
4. 發布單位：數位發展部
5. 預計發佈至官網日期：2024-1-26
6. 可參考之Figma網頁版型頁面

對齊大會成果專區 Alignment Assemblies

簡介

有鑒於人工智慧（AI）近期蓬勃發展，為確保其應用與全民利益保持一致，以進一步創造社會所需的各項應用服務，數位發展部於 2023 年 5 月正式成為國際非政府組織「集體智慧計畫」（Collective Intelligence Project, CIP）合作夥伴，參與對齊大會（Alignment Assemblies）專案，期待協助臺灣在全球公眾領域上，凝聚民眾對於人工智慧需求與風險之共識，共同處理「人工智慧對齊問題」（Alignment Problem）。

數位部以「所有人都可以決定人工智慧發展方向」為其運作目標，使全民都能對人工智慧等變革式技術的發展方向，進行有意義的參與。因此於 2023 年開始，數位部陸續舉辦與補助以臺灣使用者為核心的對齊大會，透過公民參與審議模式，形塑人工智慧發展方向，如透過點子松（Ideathon）活動，以臺灣為示範場域，並與台灣人工智慧學校基金會合作，舉辦數場活動。

人工智慧已帶來深刻的社會變革，其演算法、智慧財產權、科技倫理、公共服務和社會影響等議題備受關注，也為民主治理帶來新挑戰。因應生成式人工智慧熱潮引發社會關注相關議題，行政院正著手研擬《人工智慧基本法》草案，數位部也期待透過集體智慧計畫的先行示範，為政策制定者與技術開發者提供重要資訊，以確保人工智慧發展與群眾利益保持一致。

本頁面提供全球與國內夥伴進行審議結果分享。

目錄

1. AI 未來民主化審議式工作坊（2023/8/26及9/2）
2. AIA 人工智慧審議研討會（日期）（建構中）

AI 未來民主化審議式工作坊

1 線上審議結果

（AI Objectives Institute 共襄盛舉，運用本活動 open data 製作之視覺化互動報告）

2 現場審議結果

2.1 評估結果

臺北場綜合性建議（2023/8/26）

（一）針對 AI 對社會影響的面向，約有 3 項可聚焦處理的議題如下：

1. 勞動力釋放跟再定義：對勞動者而言，數位工具提升所釋放之勞動力，究竟是回饋至勞動者抑或是資方，將會是未來世代需面臨工作轉型的關鍵。
2. 資訊辨識與產業應用：在 AI 生成技術快速更迭的情況下，如何正確辨識 AI 所提供的資訊，將會是未來最重要的課題。尤其是產業運用 AI 來增進工作效能時，該如何避免受到 AI 的錯假訊息影響，是更為重要的議題。
3. AI 輔助或是對認知、人類創意的侷限性：如今 AI 運用將常態化，如何確保良善社會發展，保留人類生活之福祉與人性之溫暖，將會是未來社會的重要課題。

（二）在公共治理與公民參與的配套作法：

1. 建議透過例行性的公聽會，或者是民間收集意見，包含收集使用者跟廠商的意見、困境及需求，由民間提出問題，儘可能地由政府落實並制度化。
2. 建議政府透過辦理工作坊與民間互動，如這次的審議工作坊，廣泛收集大家的意見，並進一步思考，當 AI 發展落地之後，會有什麼樣的臺灣在地化需求。
3. 建議政府可透過行政指導告訴產業或使用者，溯源機制如何建立、溯源形式可行做法、可選擇方向為何，以及可能的負面影響。
4. 相對地，廠商也可以建立一種私人的反應平台來收集意見，某種程度亦為廠商相互協力的機制。
5. 最後為督促各廠商內部組成 AI 審查委員會，或組織類似事實查核團隊的工具，某種程度可建立此類民間監督的能量，同時在過程中，建立社群也可達到教育培力的目的。

（三）AI 導入未來共創方案：

1. 資料開放及管制邊界：多數組別認為需要溯源，且要載明清楚；少數組別則認為完全不需要。在轉型過程中針對管制邊界，部分組別期可補充較為嚴格的標準，包含去個資化、退出機制等機制設計，透過此類多層次的管理，鼓勵使用多元公開資料的方向，鼓勵公眾多加使用開放性資料。
2. 標準流程及競賽事項：多數組別認為應建立一個協作機制，例如透過公民參與，達成政府想要推動的目標，同時也做到教育培力，例如以競賽促進討論，並辦理類似的公眾參與活動。至於競賽項目是否分為 AI 或非 AI 組，多數組別認為尚無必要區分，但必須載明製作紀錄及過程，或是使用 AI 的比例；少數組別則認為還是有需要區分，也需再進一步思考查核技術正否有相關規範。
3. 輔導機制與擴大應用：多數組別認為 AI 可當評審或教練，尤其在教練的部分，但是做決策仍需納入人類決定的比例。在教育普及與擴大應用部分，則可使用 AI 來處理，讓這個工具可以普及化，同時要考慮到如何避免數位落差及知識落差帶來的影響。

臺南場綜合性建議（2023/9/2）

（一）針對 AI 對社會影響的面向，約有 3 項可聚焦處理的議題如下：

1. 在個人層次部分：多數組別認為資料正確性的辨識能力，將是下個世代重要的關鍵技能。
2. 在利害關係社群及社會影響部分：多數組別認為產業升級是關鍵，包含個人勞動力至整體產業如何提升，將會變成未來世代的重點與競爭力的來源。
3. 在國家應該層次上：用 AI 管制政策制度化，多數組別認為國家要設計制度，使各個場域皆可被運用，如此可推動、促進產業的升級，及回應社會需求。

（二）在公共治理與公民參與的配套作法：

1. 資料開放與監管：政府有無機會做較深度之監管，有一較為特殊之建議，即應鼓勵產業提供資料庫給政府做建檔，可加入聯合國指標的規範，或國內類似數位中間法或其他的法案等，此亦為大家所期待的議題。
2. 查核與溯源：建議強制標記產出作品的識別碼，包含提供種子碼，或倘使用 AI，則標識讓大家知道其為使用 AI 生成的作品。另一種方式為政府利用該管制工具進行查核，抑或由社群協力來布建查核的網路。
3. 教育輔導與 AI 素養課程：這是由上而下、由下而上，大家都期待要做的部分。另外一部分則為資訊素養養成與資料可信性，大家都認為可用教育普及或推廣素養教育之方式進行回應及提升品質。

（三）AI 導入未來共創方案：

1. 利用 AI 識別 AI：有關識別 AI 能力的關鍵，倘無法識別產品，有無使用 AI 或是相關分類就無其意義，但倘做好分類的標準，某種程度也是讓民眾瞭解 AI 應用具有多種層次，亦可讓這個分類更有比較的基礎。
2. 資料開放及統一規範：在統一格式碼部分，亦可連結到提高應用端的普及、AI素養教育等。從社會事件出發則希望可提出規範，再進行犯罪查核與資安防護等，此應從民間出發來提出需求，再思考如何避免並作出規範及準則。
3. 標準流程及競賽事項：有關標準及競賽事項之關鍵，則是「標的」為何，比的是技術還是人文素養，抑或人機平衡。再者，要分為 AI 還是沒有 AI，抑或比例制，例如人類跟 AI 創作，應做出一個比例區別，作為競賽類別的分類。有組別提到詠唱組或者是模型（model）組等非常細緻的一些規範，都是應被進一步釐清與確定的面向。
4. 跨邊界創新協作方案：例如由民間產業出發，協助公部門建立 AI 共創平台的架構，大家認為可能為一個解方，亦期望由政府來鼓勵、積極推動。
5. 輔導機制與擴大應用：輔導產出 AI 產品，再回饋到產業跟公共的系統進行應用，也期待過程中能平衡發展，希望技術跟人文的生產同步推進，而非僅進行 AI 發展，但忽略原本傳統或人文的社會普世價值。

2.2 討論綜整結果

本次透過兩場次的公民參與討論，廣泛地徵集意見，討論具體的政策，確立 AI 對社會影響。此外，透過公民參與，包含個人及利害關係社群，能充分反映來自民間的專業意見，不僅具有民間代表性，也作為公私協力的典範，足以作為國家施政的正當性基礎及關鍵參考依據。以下將討論結果綜整、分點，並輔以圖示說明。

（一）個人層次：勞動力釋放與賦能

對個人而言，無論職業、產業類別或專業領域為何，均認為 AI 將大幅度影響工作條件及節奏，需進一步思考所謂的勞動力釋放的紅利，究竟是回饋給勞方還是資方？另一方面，在個人層次上，亦需透過不斷地學習，取得基本操作與資訊識別的能力，方能應對未來工作或生活轉型上之需求。因此，在積極賦能的層次，除了有賴於個人與社群的專業量能，更有賴於國家提供教育培力、更多元的公民參與途徑，讓個人的能力有機會提升，並回饋至社會需求。

（二）利害關係社群層次：產業升級及輔導轉型

對利害關係社群而言，在 AI 促進產業及社會轉型的過程中，部分核心的產業可獲得升級，例如 ICT 產業等，另一部分則仰賴輔導轉型，或組織內能力再提升，例如設計、翻譯或圖文等範疇。此外，多數參與者提及，可能廣泛受到影響之產業尚有法律、教育、影像工作等。未來，職場工作者的個人數位技能，或國家單位、社群建置的培訓機制，皆可能大幅影響產業的競爭力。同時，參與者認為應透過更多的公私協力、社群之力，來協助公部門進行監督制衡。

（三）國家層次：管制政策制度化

在國家治理的層次上，多數參與者認為國家管制必須具有明確性，讓創新與研發可無後顧之憂及違反法律的風險。換言之，國家的管制必須形塑制度，讓受規範者、相關產業者及一般民眾，得以確實了解其界限，並有權利對管制界限表達意見。因此，倘國家治理的工具設計越清晰，則對個人、利害關係社群，乃至總體社會的正向循環更具助益。

2.3 潛在合作方式

本次兩場 AI 未來民主化審議式工作坊，發起自點子松（Ideathon）啟動「AI 未來民主化」（Democratizing AI Futures）對話的概念，因此本次參與者也多為點子松活動可觸及之受眾，討論內容從廣泛的 AI 對社會之影響，逐步收攏聚焦至政策工具上，包含公共治理、公民參與及徵件辦法等面向。透過審議討論設計及審議團隊（facilitators）的引導與帶領，讓民眾對於 AI 對臺灣社會帶來的影響，具有初步的理解。

未來建議，關於 AI 之社會影響，可參考本次參與者名單，對於具有較多實務見解或執行經驗者，持續追縱或定期安排深度訪談。或可由數位部連結關鍵參與者，持續性地蒐集來自各行各業的執行情況，必要時可再針對特定職業或產業類別，進行公民參與為導向之焦點座談或工作坊，增強公私部門合作之連結，讓「 AI 未來民主化」可持續性地落實在有需求的場域，從政策規劃往下扎根至回應社會需求的層次。

2.4 成果報告書與相關連結

1. Polis
2. Talk to the City
3. Talk to the City report

3 摘要

3.1 流程總覽與成果摘要

簡介：2023 年 7 月以點子松（Ideathon）為平台，啟動「AI 未來民主化」（Democratizing AI Futures）對話，並於同年 8 月及 9 月分別於臺北及臺南舉辦一場 AI 未來民主化審議式工作坊，邀請關心 AI 議題特定公眾與專家學者共同討論生成式人工智慧對社會影響的各種面向，透過持續不斷交流，思考更好的因應之道。

3.2 審議執行團隊背景

國立政治大學永續創新民主研究中心接受數位部的委託，執行兩場 AI 未來民主化審議式工作坊，該中心由一群來自公部門、實務界、學術場域所組成的團隊，致力於研發更有品質的民主參與程序及社會對話方法。該中心想要實踐的「社會對話」，是從「議題」出發，與社會夥伴進行各種形式的溝通和合作。具體來說，是由與政策議題相關利益的社會夥伴包含公民行動者、組織團體、專家學者、政府部門等，進行跨領域或是跨層級的對話，以討論、協商、諮詢、共同行動等方式來面對當代複雜的問題，建立共識以及促進民主參與。

「永續創新民主研究中心」以創新思維開拓公共參與之於公共政策的管道，該中心近期將研究焦點放在科技治理之面向，包含科技鄰避設施與人權、科技創新對社會影響評估、核廢料選址與公民參與等，試圖將公共議題置於檯面，期望能藉由多方論壇、意見整合後得出最適合社會發展的解方。

3.3 執行動機與問題意識

生成式人工智慧具備大數據運算、即時產製內容的特性，依賴機器深度學習技術的廣泛運用，從而發展出具備仿效、創造的能力。人工智慧生成內容（AIGC）的應用，有機會改變現代社會的勞動結構、產業布局，甚至是應用在日常生活習慣上，對社會運作帶來不同程度的挑戰與機會。根據全球最大程式碼管理平台 GitHub 針對美國程式開發人員之調查，有 92% 的受訪者表示他們在工作上運用生成式人工智慧撰寫程式，並有 70% 的受訪者認為人工智慧的運用，能提升程式編寫的品質與工作效率；另一方面，過去多數人在創意創造力產業，將人類智慧財產視為值得以法律保護的主體，現在若透過生成式人工智慧結合創意所生成的作品，其效率及精準度表現皆十分優異，而其中的智慧財產歸屬恐掀起另一波討論；再進一步，若談論到對社會的負面影響，倘若人們本身帶有的社會觀點與偏見，透過生成式人工智慧生成帶有偏見的內容，再透過傳播系統的快速傳散，某程度也將加快歧視思維傳播的效率，並深化社會對立，皆是在發展與使用生成式人工智慧必須審慎評估的部分。

新興科技的發展十分快速，尤其是生成式人工智慧已成為國際各大科技產業競逐的新興舞台，生成式人工智慧的發展為社會揭開了新秩序的篇章，國際間也針對該議題開始思辯與提出管制方案，例如經濟合作暨發展組織（OECD）便將 AI 界定為「以機器為基底所設計的系統，可以預測、提供建議或決策，來影響實體或虛擬環境，達到人類設定的目標。這個系統具備某種程度的自主性，而且是可以執行類似人類認知功能的機器」，並提出所謂的《OECD AI 原則》（OECD，2023年：17），其中包含五項對國家政策和國際合作的建議：（1）投資於 AI 的研究和開發；（2）培育 AI 的數字生態系統；（3）塑造有利於 AI 的政策環境；（4）建立人力資源並為勞動力市場的轉型做好準備；（5）進行可信 AI 的國際合作。在 OECD 的研究報告中也指出，生成式人工智慧將篩選大量數據，可用於幫助政府制定政策決策，其可預測和幫助防止選舉干預事件，改善公共服務等優點，也認同生成式人工智慧應用於政府治理，惟提醒生成式人工智慧可能通過延續和放大社會不平等來削弱民主價值觀，若不予以適當管制，則可能進一步破壞信任和社會契約。

回到我國的脈絡，生成式人工智慧對社會影響如國際趨勢，與民主、社會與政治的交互關係顯而易見，但現階段我國多數的討論仍聚焦在產業發展居多，例如評估私部門應用 AI 為企業節省成本或創造收益等，然而在開發與經濟導向之外，進一步思考生成式人工智慧對民主、社會與政治的影響面向又有哪些，便是對於政府治理與社會永續發展皆十分重要的提問。觀諸學術研究發展方向，近期有不少研究開始對生成式人工智慧可能為公部門帶來的諸種倫理威脅產生疑慮，包括不同階段的科技浪潮，將可能對政府帶來不同面向的轉型，但也可能創造新的公共服務與價值。另一方面，針對社會個體或不同的產業類型，又有哪些較受關注的議題值得深入研究，則需要有更多的實證資料來佐證，引發更具有政策回應性與社會需求導向的研究方向。

據此，國立政治大學接受數位部的委託，從人文社會觀點切入，結合數位部 Polis.tw 民意調查報告，蒐集對於 AI 議題具有想法及有討論意願的民眾，評估議題特定公眾對於生成式人工智慧較具分歧的想法為何，透過討論發展出具有臺灣脈絡及臺灣社會需求的爭點，以利主責機關未來有機會可以透過政策工具來解決或改善可能的風險。

3.4 執行限制

（一）參與者代表性界定

本次兩場 AI 未來民主化審議式工作坊，由於發起自點子松（Ideathon）啟動「AI 未來民主化」（Democratizing AI Futures）對話的概念，因此本次活動參與者多數是來自參與點子松活動且有意願參與工作坊的投件者，另一部分則來自受邀參與之專家學者。換言之，本次參與者來源可謂是採特定對象群體之「自願制」及「邀請制」雙軸並進組成，因此活動參與者並非全部都是第一線使用者，也可以看出來自設計領域的比例明顯高於其他領域，各自對於技術使用的了解、想法及表現也有所落差。當然參與者邀請的來源、數量或代表性之適當性，取決於活動設計的目的性，因本次兩場討論活動欲回應的主題之一，係因應生成式人工智慧興起，點子松徵件辦法之修訂方向，因此本次的參與者組成確實連結及扣合討論目的；對於活動議程想要以更廣泛地討論 AI 對社會影響來進行設定，此一方向仍可透過不同職業類別的參與者提供觀點取得初步的藍圖。未來倘有擴大辦理類此議題討論之規劃，則可在參與者代表性上多加設計，包含可採分眾分群辦理，部分場次邀請制、部分場次自願參與制，以符合活動目的及協助主辦方取得必要的意見來源。

（二）討論議題範疇主題化

本次兩場 AI 未來民主化審議式工作坊，討論的題目從民眾意見開始蒐集，並透過 Polis.tw 歸納意見內涵，再輔以國內外新聞、研究趨勢等面向，收攏至主題（包含政府治理-內部應用、新興法治-監管機制、資訊識別、資料開放、智慧財產、教育與人類認知影響）進行討論。因參與者多數為有意願參與工作坊之點子松投件者、部分為受邀參與之專家學者，爰執行團隊將專家學者分配至與其專業領域相近的組別。而有意願參與的投件者因來源較為多元，基本上以隨機分配方式進入議題組別。因此，各組內意見表現水準不一，有來自學術層次、實務經驗或是出於想像者，參與者不見得對於被分配到的討論主題能夠全盤掌握。然而，事實上本次討論的 AI 應用技術尚未全面性地普遍或成熟，多數人對於 AI 在實務運作的限制與困境，還未能通透地了解或具有足夠經驗累積。換言之，AI 應用所產生的問題，並非是具有公民共識即可操作的知識能力，反而是需要真實具有接觸、訓練或使用經驗者，才能夠更加深入分析、提出具有前瞻性的建議。因此倘若未來還有機會再辦理類似的議題討論，可先聚焦討論議題，選出較單一的主題或對應到特定職業或產業類別，進行較深入的研討，方能夠對於 AI 對社會影響的面貌與實務應用，具有更精準的掌握。

（三）政策建議追蹤與落實

由於以公民參與為導向的活動設計，參與者需貢獻不少行前準備的時間，以及現場交流所負擔之心力，因此願意參與之參與者，多半會具有一定程度之參與效能感，或者可透過討論提出相關見解，以及具有被採納的預期心理。一場好的審議活動，即是透過審議團隊的事前知情及現場帶領，聚焦眾人意見，產出參與者感興趣且同意的結論，進而產出有效的政策建議。然而，雖然現場討論熱絡、參與效能感極高，在後續是否真的能夠推進實際政策議程，或結論是否具備一定的政策影響力，仍需回到具有權責的公務機關，看其履行的程度來確定。根據往例，類似的討論活動結束後，參與者不容易有心力持續追蹤後續的政策落實，而政策規劃與推動亦需要時間，較難立即見影，這樣具有政策導向或社會發展的議題特性，也使得後續效果較難以確認。因此，在活動場上的對話結果，是否對真實政策能夠發揮影響力，亦或是討論的內容是否回應到一開始啟動討論的初衷，則仍有待透過實務工作者之運作來檢驗，這些對接政策實務的產出，確實非審議活動當下能夠確定或保證的範疇。

4 方法

4.1 研究方法與參與成員身分分析

本次兩場次活動的參與者共計 95 位，性別比例方面，男性比例稍微多於女性，男性占比 61.1%，女性占 38.9％；以臺北或臺南兩場次獨立分析來看，性別比仍是男性多於女性，可參考下圖：

在年齡分布上，若將兩場次的參與者人數併計，年齡區間落在 18-29 歲的參與者為多數，占比大約 4 成，接著是 30-39 歲，占比 24.7%，再來是 40-49 歲，占比 22.4%，由此可知，參與者年齡低於 49 歲者，占比高達 87% 以上，顯見本次討論年齡多數是落在青年、青壯年世代的參與者，可參考下圖：

在參與者所屬單位類型的分析上，若將兩場次的參與者人數併計，來自學校或政府智庫類型的比例大約 41.9%，跟來自民間團體或公司行號的比例接近約 40.9%，其餘則是來自政府單位。由此可知，本次參與者來自學術單位跟來自第二及第三部門約為各半，來自第一部門大約 17.2% 左右，可參考下圖：

在參與者專業領域分布的分析上，若將兩場次的參與者人數併計，來自數位科技與設計領域者超過 6 成，其餘還有來自社會科學、法律、商業、醫學、心理學及教育等領域；專業領域高度集中在數位科技及設計領域，其組成與本次參與者之邀請主要為點子松投件者具高度相關，可參考下圖：

4.2 流程細節（議題框架與討論設計）

政治大學創新民主中心所提倡的社會對話方法，是從議題分析出發，盤點議題現況所可能帶來的挑戰、待解決的問題，及利害關係人之辨識，思考在此議題脈絡下，利害關係人等各個角色在此議題之目標為何。再者，可視需求進入議題擴散階段，透過不同層次的對話框架設計，客製化符合實務需求的操作方法。同時該中心所擅長的社會對話設計，是在維持審議思辨的同時，同等強調換位思考、設身處地的重要性，期待每次的討論不只是停留在論理，更能夠透過身體感知，將議題關懷帶入日常生活。

本次工作坊依循政治大學創新民主中心工作方法，透過對於生成式人工智慧議題現況的盤點與分析，再搭配 Polis.tw 的意見蒐集，辨識出目前社會公眾在思考生成式人工智慧會設想到的情境。最後，再透過兩場次審議式工作坊的活動設計與操作，探尋不同議題爭點的成因，以作為後續政策規劃的參考。以下簡要說明本次活動設計概念與操作方案：

（一）情境設定

目的是在活動開場時，透過行動劇的展演，讓參與者較容易且無進入門檻地了解接下來的議題討論目標。以下為行動劇的情境設定：

在生成式人工智慧的多元宇宙世界，究竟會如何改變我們現在的生活型態呢？

場景-某公司

A 擁有一間「協助產業開發生成式人工智慧」的公司。近年正逢生成式人工智慧大規模發展，再加上生成式人工智慧開發專業性高，資源一把抓的 A，資本也在近幾年大幅提升。最近他正協助 B 公司進行設計圖像機器人研發，B 公司的員工 C 非常開心有生成式人工智慧可協助創意生成，讓他的工作效率大幅提升，改善了過去需要過勞加班的黑暗期。

有天員工 D 在公司午餐後閒晃，看到一個競圖比賽徵件，發現離截稿時間只剩下不到 10 小時，看到誘人的獎金實在很想參加，但工作又很忙實在沒時間製作，於是靈機一動想到可以應用生成式人工智慧協助生成圖像再來修改就好，果不其然只花不到 6 小時就把圖改好投件了。

不料 D 的提案仍被審稿人 E 所識破，認為有生成式人工智慧的協作，這個作品不能夠算是 D 全權所有的著作，E 甚至進一步發現 D 所生成的內容，部分是來自於他人的著作，卻未能有清楚註明資料來源，恐違反作品發表的倫理，因此向主辦方提出應不予參與的異議。

場景-學校

在多元宇宙的中學裡，老師利用生成式人工智慧作為授課的素材與教學工具，學生們也很習慣應用生成式人工智慧來學習及完成作業。像是 F 學生就表示每次遇到想不通的題目，就可以打開 ChatGPT，一下子就完成了呢!然而，對於 G 老師而言，如何避免學生過度地使用生成式人工智慧來學習，變成在多元宇宙的老師們共同的難題。

有天 F 學生在課堂上發表他透過生成式人工智慧撰寫的程式碼，可以將老師的照片改得像是林志玲， F 學生非常得意地將這套軟體開發得更容易使用並分享給同學們。有天 H 同學閒來無事便拿著這套軟體開始改圖，以好玩的心態製作多種素材，其中有一款是拿 F 同學的人像製作成哭泣的畫面，不料卻被同學拿來傳散，傳到了 F 家長 I 的手機裡，引發軒然大波。

甚至 I 家長還跑到學校教務處申訴，以為 F 被同學欺負了 I 家長走到教務處希望教務長 J 可以好好整頓學校風氣，這時便看到教務長 J 正利用生成式人工智慧產製公文書，多元宇宙的公務系統早已研發出一套生成對抗網絡，可查核申請資料填寫是否正確，透過技術協助十分快速地謄寫完畢。J 聽完事件始末後，得知是學生之間利用生成式智慧技術玩得太過火了，便開始思考如何避免學生對事件的認知脫離現實，或許也該重新整頓教學模式了。

（二）分桌與議題設定

在行動劇展演完畢後，本次討論預計分六桌，討論主題為政府治理（內部應用）、新興法治（監管機制）、資訊識別、資料開放、智慧財產、教育與人類認知影響等，各桌提問流程將先定義出各個議題面向的挑戰與爭點，再以探討何種情境或條件下，可提高人類對於生成式人工智慧應用的親近性與信任度，透過討論辨識出優勢、挑戰與風險後，進一步思考可能的解方。各桌討論的爭點說明如下表：

（三）討論模式及工具單設計

根據上述規劃之議題框架及討論構想，以結合議題知情、同理與感受等元素進行討論流程之規劃與設計，同時考量參與群眾多數為對 AI 議題有所涉獵之專家或感興趣的相關社群，多數參與者對於議題具有專業研究或經驗性見解，相較於一般公眾的討論模式而言，本次討論將更加聚焦及具有可操作性。

因此本次將討論模式定性為以特定公眾為主之開放式焦點座談，相較於不設限議題爭點的討論模式，傾向由各組別針對不同主題進行個案研討，再由桌長進行觀點收斂與建立討論流程之延續性，以利同時符合參與者對議題掌握的調性，同時又可透過參與者提供之論述收斂出對議題發展有助益之意見觀點。

本次各組別的參與者組成大致可以歸納如下圖，包含 Polis.tw 社群或資訊工具使用者、相關領域之專家學者及具公共治理視角之對話，主要考量到本次討論的終極目標為期待對公共治理有新的想像與尋求可能的共識，因此除了參與者本身特性（例如專家或社群觀點）外，特別強調須具有公共治理之視角，以利討論結果可提供後續應用，此部分的設計也將於討論工具單與桌長引領技巧中被運用及落實。

因此，本次討論依據討論主題預計分為 6 組，每桌的組成人數約為10位，相關社群代表約 6 人、專家學者約 2 人，帶領討論及引導公共治理視角介入的桌長群為 2 人，單一場次大約 60 位觀點貢獻者，如以下圖示：

本次活動分組討論架構分為兩大部分，第一輪針對「 AI 對社會影響之現況盤點、優勢與挑戰」、第二輪針對「如何透過政策工具促進 AI 與人類社會創造共融新生活」，在審議設計層次上，第一輪主要是透過拋出問題，引發參與者思考自身與議題間之連結，也是討論流程中所謂的「暖身」，在拋問的過程中，協助各組內的參與者建立互相理解的機會，有助於每一組就分配主題進一步研討與分析。每一輪討論尾聲，藉由小組分享讓參與者了解跨組別意見，同時對於議題建構具有更完整的想像。

第二輪則是以上午跨組別分享後的觀點，由桌長引導以公共政策形成最基礎的分類，分別是由上而下的公共治理與由下而上的公民參與模式，進一步思考如何透過政策工具回應上午各組所提出的風險與挑戰，透過排序與條件互相搭配，發展出兼具可執行性與創新的行動方案，來共創人類與 AI 共融的新生活。

本次亦綜合活動目的與討論規劃，設計有助於現場討論推進的工具單。首先，在創造個人與議題連結的層次，以正向與負向兩個指標，引導參與者分享自己對 AI 運用的觀點，同時透過紀錄單所留下的紀錄與歸納，讓參與者之間有機會互相觀看，了解自己組內參與者對於 AI 的偏好與應用邊界。

接著進入議題研討階段，主要聚焦在 AI 針對個別面向之影響（包含優勢與挑戰），此部分應用社會結構分類框架，以個人認知、利害關係社群、社會影響與國家治理等構面，從個體到總體的思考脈絡，引導參與者分享自己對於議題觀點與更深入的判斷，進一步推估對於社會不同層次的單元會有哪些不一樣的需求與挑戰，這個階段不僅停留在提出問題的層次上，而是進一步引導參與者展開原本對議題思考的框架、試圖協助參與者創造對於議題更多元的想像。

在第二輪討論部分，終極目標為提出可能的共創方案，因此在延續上午末段各組分享對於議題的理解及諸多有待解決的問題與挑戰後，在試圖跨到提出解方的層次前，先以拋出期待、重新拉起討論動力的設計，讓參與者理解，即使有諸多待解決的問題與挑戰，仍可進一步思考為何人類仍希望創造與 AI 共存的未來社會，而要達成這個理想，現階段又有哪些政策工具可供使用，在此層次上試圖再次分層，將政策工具分為法令/規則、策略/方案及競賽項目等三大項次，引導參與者將解方略為分類，以利後續落實到政策應用時，可有效對接至相應之政策發展階段。

在實作前，透過桌長工作坊進行內部演練，以確立設計框架及拋問爭點符合原設計的想像，且可回應業務單位的期待。透過演練，也確保框架可由桌長落實，並連結議題與實務。現場討論操作時，主要由桌長群（facilitators）依步驟引導參與者分享觀點，隨桌紀錄通常會以 N 次貼或彩色筆繪圖的方式，將意見關鍵字撰寫在海報紙上，讓參與者可來回觀看，並透過意見堆疊、自身意見觀點建構，最終由桌長進行歸納與整理，形成組內具體建議。

4.3 線上討論流程的輸入與案例

Polis 分析

在本次活動設計截止前（統計至 2023 年 8 月 1 日止）來自 Polis.tw 調查，共有 403 名民眾參與投票，共做出 8,591 次選擇，平均每一位都投出 21 次以上的投票選擇，共提出 126 種論述，共有 260 人被分組。以下簡要說明本次 Polis.tw 民意調查對於生成式人工智慧應用的意見偏好：

（一）同意比例高者意見分析

大致上可以將民眾具有共識之陳述內容分為兩大類，其一是比較正向且樂觀地看待 AI 服務，認為其僅是工具性、輔助性或協作性質的產品，無需過度擔心，認同可用以提升人類工作的效率，至於利弊與否需端視人類如何使用；另一種則是認為 AI 恐會改變部分社會現況，並提出在部分管制界限上的疑慮。

編號	內容	同意比例
118	AI 生成的內容應適當標註參考的資料來源。	95% 同意（N=145）
44	隨著人工智慧技術不斷向邊界的速度發展，人類與機器生成的創作之間的界限日益模糊，關於知識產權、知識產權和法律責任的關鍵問題需要被深入探討。	94% 同意（N=155）
54	公務人員使用 AI 處理業務內容時，應進行相關培訓。	92% 同意（N=166）
104	在安全防護與審判上，不能全部依賴 AI，人性也必須考慮。人與 AI 最大的差異正在於此。	92% 同意（N=157）
45	目前人工智慧的發展狀態，將會令人驚嘆地改變許多產業生態。	91% 同意（N=155）
48	關於AI的規範，要考量到世界各地不同的文化和脈絡，而不僅僅是特定國家的觀點。	88% 同意（N=171）
55	使用人工智慧能夠減輕國家公務員的負擔，但需要研究討論如何處理機密訊息時，不會發生訊息外洩的疑慮。	87% 同意（N=165）

（二）較具有共識者意見分析

從本次報告中可以看出分成三組，B 組可看似為比較中性意見的組別，而 C 組是對於 AI 工具較為樂觀看待，則 A 組可視為比較重視人與人互動，認為 AI 將部分改變現況的組別。在分析同意比例高者意見後，可以參考分組成員皆認同，也就是具有跨組別共識的意見有哪些，一方面可與同意比例高者的意見進行核對，另一方面可作為後續選擇討論議題的重要參考。目前以共識程度高到低來觀察，仍不脫離上述分析的框架，但可以觀察到一個現象，樂見其成與認為會部分改變現況的意見多數無互斥現象，且部份是具有跨組共識的論述，以下將具有群體共識的意見列舉如下表：

編號	內容	同意比例
44	隨著人工智慧技術不斷向邊界的速度發展，人類與機器生成的創作之間的界限日益模糊，關於知識產權、知識產權和法律責任的關鍵問題需要被深入探討。	94% 同意（N=155）
80	以正向態度運用，學會共創善用。	94% 同意（N=169）
85	AI 的應用還算青少年，是機會還是威脅，還有很大的探索空間，不用劃地自限，但要預作因應。	95％ 同意（N=185）
117	人工智慧是一個工具，但所提供的訊息資料，假真真，共謀真偽，才是個人專業的表現，共利用的人將會提高工作的效率，類似於Google的崛起，人工智慧的崛起，又是另一種思維模式轉變。	100% 同意（N=27）
45	目前人工智慧的發展狀態，將會令人驚嘆地改變許多產業生態。	91% 同意（N=155）
104	在安全防護與審判上，不能全部依賴 AI，人性也必須考慮。人與AI最大的差異正在於此。	89% 同意（N=77）
83	AI 的東西，其底部判斷標準還是回歸到人。	92% 同意（N=152）
54	公務人員使用 AI 處理業務內容時，應進行相關培訓。	92% 同意（N=155）
106	AI 不能取代所有的生活技能，但AI可以是生活更便捷。	94% 同意（N=37）

（三）意見較分歧面向
針對參與意見調查者對於單一論述提出同意與不同意比例高度分歧的意見，有機會為現行社會脈絡下特別需要提出來討論的爭點，也是後續在規劃討論議題分析的重要參考來源，因此針對意見較分歧的面向列表如下：

編號	內容	同意比例
56	對於政府使用 AI 進行業務處理，是讓人信賴的。	36% 同意；35% 不同意（N=48）
47	當未來人工智慧充滿真實生活時，整個世界從目前的後真相時代進一步成為無真相時代，完全放棄或無能追求。	42% 同意；36% 不同意（N=183）
10	新科技的恐懼被過度放大，人類其實並不需要理解背後是如何運作的，也能夠利用它並完成批次。	55% 同意；34% 不同意（N=185）
66	過度依賴人工智慧可能會導致學生缺乏人際交流的機會，影響他們的社交能力。	53% 同意；25% 不同意（N=189）
60	AI 將原始數據破壞壓縮，學習過程會更難獲得真實數據。	45% 同意；27% 不同意（N=156）
39	由人類自己創造的東西是（比）人工智慧創作的，更有意義。	60% 同意；20% 不同意（N=163）
69	我認為投稿作品可以在 AI 生成後再使用 Photoshop 製作。	66% 同意；15% 不同意（N=178）
6	隨著創作量大、質量高的衍生作品的越來越低，版權成本將越來越不受約束。	64% 同意；22% 不同意（N=166）
67	我覺得 AI 作品應該單獨設立一個獎項。	66% 同意；15% 不同意（N=48）
64	身為教育者，需要提供更貼近生活、親身感受式的經驗，而不是 AI 生成的內容。	63% 同意；13% 不同意（N=165）

（四）建議討論議題面向

依據上開初步分析，從多數人同意的選項中，可以識別參與投票的民眾對於生成式人工智慧的態度，約可分為兩大項，包含樂見其成與認為需要部份管制的觀點，進一步在跨組具共識的分類下，仍可見此兩大類的論述並無明顯衝突，多數意見是跨組別仍認同的論述內容。因此至少可推論，以本次填寫的母體範圍而言，或許「對於生成式人工智慧正面看待，但部分業務認為需要提出管制方案」這樣的看法，可暫時視為是本次調查的多數意見。

再進一步針對意見分歧的部分來看，則可進一步排序出引發使用者困惑或有疑慮的面向有哪些，包含政府治理（分有內部應用、監管機制）、資訊識別、資料開放、智慧財產、教育與人類認知影響等，這些議題面向大概是對於生成式人工智慧抱持著正面看待，但仍認為部分議題需要進行管制的爭點，因上開列表議題，同一群人在選擇上出現明顯的意見相左，背後成因則有待進一步透過質性資料的蒐集來釐清。

因此，觀諸國際趨勢、國內議題現況，以及本次 Polis.tw 的意見搜集，建議本次議題討論工作坊就以政府治理（聚焦內部應用）、新興法治（聚焦監管機制）、資訊識別、資料開放、智慧財產、教育與人類認知影響等六大主題進行討論題目設計。

以上分析結束後，相關問題交付給現場審議工作坊民眾討論，並由專業引導師收攏。

4.4 方法論設計與實作

在民主社會，透過直接民主（direct democracy）、代議民主（representative democracy）等機制運作，以選票及定期選舉選出代議士，代表人民行使公權力的民主模式實屬常態。然而透過選票加總式、競爭性的民主運作模式，或許可達成國家治理必要之運作有效性及決策效率等目標，卻仍非是毫無缺點的運作模式，通常在面對多數決策、贏者全拿的選舉慣性下，若是屬於少數公民，則容易覺得自己的意見難以被充分表達及回應，進而有所謂多數暴力的風險。當代學者為了彌補加總式民主的挑戰，因而提出所謂審議式民主（deliberative democracy），主張讓公民透過公共審議的方法參與公共事務的討論與形成影響決策的集體意見。

同時，隨著當代社會各方面突飛猛進地發展，不論是公共議題越趨複雜且跨領域，加上公眾對於資訊公開的掌握及需求越趨提升，便不難觀察到許多時候政策規劃與推動，不再只是在尋求政府與社會之間的妥協或是共識的層次，更多時候需要更積極面對來自社會群體之間的多元差異。因此現代社會亟需有一個疏通、互動的管道，讓更多的討論及不同意見可以被呈現出來，甚至是透過互相觀看、理解的過程，形成新的共識或集體意見，這樣的概念便與所謂的審議式民主不謀而合，良好的社會對話實踐與落地，將影響社會對於民主深化的理解，並有助於提升公共討論的品質。

因此，本次兩場次的活動設計之所以稱為審議式工作坊，便是大量參考審議式民主的精神，在公共政策規劃與設計的過程中，透過知識系統的整理與轉譯，再輔以公民參與的模式，邀請公民進入議程討論、發表見解，並透過整理、歸納出經過討論的建議方案。因時間有限性，此類繁瑣的過程並非召開會議即可自行運作，而是需要透過專業的審議團隊進場帶領，讓參與者可在最短的時間中，理解需要討論的公共議題內涵，並在有限的時間內能有一定之結論，與透過設定足夠具體的問題，達到收斂出具有建設性或執行可能性的政策建議之目的。

如前所述，本次執行團隊－政大永續創新民主研究中心－於審議操作上有諸多議題討論的經驗，強調所謂感性同理的應用，尤其是再現審議民主中，十分強調的相互性及公開性。透過他者的意見進而反思、打開自己的觀點，具備推動審議民主時的關鍵經驗。本次審議式工作坊亦結合劇本設計的成分，由團隊成員現場即興演出有關於 AI 對社會影響的短劇，讓參與者可透過劇情，發想可能的情境，並與自身經驗進行扣連，讓後續的討論不只存於論理的層次，同時也能夠更加貼近現實社會運作的樣態與需求。

4.5 方法論影響與副作用

承上所述，本次兩場次工作坊的活動設計及現場操作，皆大量採用審議式民主精神，企圖再現審議式民主模式所強調的優勢。方法上，先由執行團隊進行議題研究，製成簡單好閱讀的活動手冊，再透過設定議程及討論主題的方式，期在有限時間內有效率地讓參與者多方發言，並透過審議團隊的帶領，進行意見收攏與聚焦。同時，考量到進入議題所需的感知啟動，因此設計透過展演的方式，開啟參與者的感官，讓討論的過程較不容易流於紙上談兵，而是更貼近人類社會可能出現的情境。

這樣的操作方法自然有其正向影響，包含可以在有限的時間內，透過審議團隊具有技巧的引導與帶領，讓每個參與者都能夠發表自己的意見，並透過互相觀看的過程中進行交流、修正主觀性看法，進而尋求集體共識。同時，透過設定足夠具體且清晰的問題，便有機會蒐集到具有建設性及執行可能性的政策建議，並透過由參與者主動參與討論所延伸出的相互性、反思與接納，進而能夠對現場所提出之建議方案，產生同理與抱持著接受的態度。透過審議所達成的正向循環，不論現場所討論的議題回到決策層次，是處在政策規劃、合法化或執行階段，皆可對公共決策者在推動政策時有所助益。

反之，帶來的副作用略可分為兩大面向，其一是類似的審議討論，其活動社會動員的成本相當可觀，通常可全程參與的人數有限，加上若要考量到討論品質，單場次也不宜邀請過多參與者，通常的活動規模大約落在 20-50 人次較為常見。因此，除非有機會可舉辦多個場次，否則在單場次可參與人數之限制下，或許仍難以收納到全面性的意見。倘某一場審議式民主活動，對於回推至社會代表性有高度且嚴謹之要求，也可考慮採用系統性抽樣，或立意抽樣的方式來邀請參與者，以避免參與者之組成偏頗，影響意見代表性的風險。

其二則是會議結論需經過驗證的副作用。由於審議式民主活動的辦理成本較高，因此無論是主辦單位或是參與者，對於結論的可執行性與影響力將具有較高的期待，然倘為政策議題的討論，相關結論仍需回到權責機關予以落實，方能檢驗討論結果，在未來執行之可行性。直至推動的當下，或許才有機會檢驗討論現場所推論方式之正確性。此為審議式民主在討論過程中，無論討論氣氛或討論結果為何，皆難以主張討論當下之集體共識及會議結論，一定能夠在政策設計上被有效地落實。

因此若要改善上述所謂的副作用或風險，即為主辦單位意識到預計開啟討論的題目，並非是一般性社會問題，而是可能即將形成政策議程的爭點，便需要在參與者邀請時，確保受邀的參加者能夠針對議程發表洞見。同時，需要由審議團隊擔任中介角色，進行議題資料梳理及歸納，提出具有討論價值的議程設定，討論時廣納、平衡參與者的多元意見，並依據會議目的，收斂出具體且具有建設性的建議方案，以利具有審議式民主價值的公眾討論，能夠對接到政策實務、發揮影響力。

---

<以下為英文版>

Introduction

In May 2023, the Ministry of Digital Affairs (moda) partnered with the international nongovernmental organization Collective Intelligence Project (CIP) to participate in the Alignment Assemblies project to ensure the application of artificial intelligence (AI) is consistent with the people’s interests and create various application services essential to society. The goal is to help Taiwan reach a public consensus regarding the needs and risks of AI in the global public sphere and collaborate to resolve the alignment problem of AI. Recent developments in AI have prompted this partnership, and the moda is committed to aligning the use of AI with the people’s interests while creating essential application services for society.

The operational target of the moda is for everyone to determine the development direction of AI so that all people can meaningfully participate in the development direction of transformative technologies such as AI. Therefore, the moda has organized and supported the Taiwan-user-based Alignment Assemblies since 2023 to shape the direction of AI development through citizen participation and deliberation models. One example is the Ideathon event, which uses Taiwan as a demonstration site and collaborates with the Taiwan Artificial Intelligence Academy Foundation to organize various activities.

AI has ushered in profound social changes. Issues such as AI algorithms, intellectual property rights, science and technology ethics, public services and social impact have attracted much attention. AI has also introduced new challenges to democratic governance. The Executive Yuan is drafting the “Artificial Intelligence Basic Act” in response to the rise of generative AI, which has raised social concerns about various issues. The moda also anticipates providing critical information to policymakers and technology developers through the CIP pilot demonstration to ensure that AI development is consistent with the interests of the people.

This page shares the review results with both global and domestic partners.

Table of contents

1. Democratizing AI Futures Deliberative Workshops (2023/8/26 & 9/2)
2. AIA Artificial Intelligence Deliberation Seminar (date) (under construction)

Democratizing AI futures deliberative workshops

1. Online deliberation results

(AI Objectives Institute participates in this event and uses the open data from this event to produce a visual interactive report.)

2. On-site deliberation results

2.1 Evaluation results

Recommendations from the Taipei session (2023/8/26)

(I) There are essentially 3 concerns to focus on regarding AI’s impacts on society:

1. Labor force released and redefined： Whether the labor force released by technological advancements is returned to workers or employers will be critical to the work transformation that future generations of workers will face.
2. Information identification and industrial application： Due to the rapid changes in AI generation technology, correctly identifying the information provided by AI has become the most important issue in the future, particularly for industries that use AI to improve work efficiency. However, how to prevent the impact of false information provided by AI is even more critical.
3. AI assistance may have cognitive and creative limitations： As AI applications become more prevalent, a critical issue in future society will be to ensure development while preserving human connectivity and well-being.

(II) Complementary approaches to public governance and citizen participation:

1. It is advised to hold regular public hearings to consult with the private sector and gather information regarding manufacturers’ and users’ needs, concerns and opinions. The private sector can raise issues to help facilitate systemic implementations by the government.
2. It is recommended the government organize workshops and interact with the people. Similar to this deliberative workshop, the goal is to gather more feedback from everyone and consider what needs will arise for Taiwan local once AI development has been implemented.
3. It is recommended that the government should provide information to private sector through administrative guidance or show users how to trace the sources. What can be done in terms of traceability, what are the possible directions and what are the potential negative consequences?
4. Manufacturers can also establish private response platforms to collect feedback. To a certain extent, this can also become a mechanism for manufacturers to collaborate with one another.
5. Finally, we encourage manufacturers across industries to form internal self-review committees for AI or organize tools akin to fact-checking teams. This type of private supervision energy can be established, and building such a community can also facilitate education and empowerment.

(III) Integrate AI into future co-creation plans:

1. Data openness and control boundaries: Most groups believe traceability is necessary and must be clearly defined, while a few feel it is unnecessary. Some hope that de-personalization, exit mechanisms and other more stringent standards for controlling boundaries should be added during the transformation process. Through this multi-level management, we can encourage public data diversification and inspire the people to use more open data.
2. Standard process and competition-related matters： Most groups believe that collaborative mechanisms such as citizen participation should be established to fulfill the government’s goals while achieving education and empowerment. For example, most groups are willing to organize public participation activities such as competitions to promote discussions. As for whether competition events should be divided into AI and non-AI groups, most of those present believe there is no need to make such a distinction. However, the production records, processes or the proportion of AI used must be clearly defined. Of course, few groups believe such distinction is still necessary, and they should further inspect whether there are any relevant specifications for the technology.
3. Counseling mechanism and expanded application： Most groups believe AI can be used as a judge or coach, particularly in coaching. But decision-making must still include a proportion of human decisions. AI can be used to popularize education and expand its applications. Meanwhile, consideration must be given to avoiding the impact of digital and knowledge gaps.

Recommendations from the Tainan session (2023/9/2)

(I) There are essentially 3 concerns to focus on regarding the impact of AI on society:

1. At the individual level, assessing data accuracy across multiple sources is a critical skill for the next generation.
2. Most groups believe industrial upgrading is critical at the stakeholder community and social impact levels. The key to future competitiveness is enhancing individual workers and the overall industry.
3. At the national level, most groups believe that the country should use AI regulatory policies to institutionalize a system that can be employed in various fields to drive and promote industrial upgrading and respond to social needs.

(II) Complementary approaches to public governance and citizen participation:

1. Data openness and supervision： This session made a rather particular recommendation about whether the government should conduct more in-depth oversight. The private sector should be encouraged to provide databases for the government to build records. All parties look forward to including the United Nations indicator specifications in domestic legislation similar to the Digital Intermediary Act.
2. Verification and traceability： It is strongly advised to compulsorily mark the finished product’s identification code or provide a seed code. If AI is used, provide a mark to inform everyone that the work was generated using AI. Another method is for the government to use this control tool for verification or for the community to collaborate and build a verification network.
3. Educational counseling and AI literacy courses： This is the top-down, bottom-up part anticipated by everyone. The other part is to improve information literacy and credibility. Everyone believes that we can respond and improve quality by popularizing education or promoting education literacy.

(III) Integrate AI into future co-creation Plans:

1. Use AI to identify AI： The key to discussing identification capabilities is that using AI or related classifications will be meaningless if we cannot identify the products. However, creating classification standards will also help inform the people about the various levels of AI applications. Secondly, these classification standards can also be used as a basis for comparison.
2. Data openness and unified standards： The unified format code component can also be linked to the need to increase application popularity, AI literacy education and other areas. From the social events standpoint, we hope to propose regulations and conduct crime audits and information security protection. We should begin by considering the people’s demands and then determine how to implement prevention while establishing norms and standards.
3. Standard process and competition matters：Determioning what the objective is in terms of standard and competition is critical. It determines whether the competition is based on technology, humanistic traits or human-machine balance. It is necessary to differentiate if AI, no AI or a proportional system such as human and AI collaborative creations is used. A proportional distinction should also be made to classify the competition categories. Some groups were classified as singing groups or model groups based on particular criteria, and all of these criteria should be further clarified and determined.
4. Interdisciplinary innovation collaboration projects： Everyone believes that starting with the private sector and assisting the public sector in establishing an AI co-creation platform structure could be a solution, and we hope the government can encourage and promote such projects.
5. Counseling mechanisms and expanded applications： Provide counseling on AI product development, which will be fed back to industry and public systems for applications. We should aim for a balanced development of technology and humanities while also preserving traditional humanistic values rather than solely focusing on AI development and ignoring the original, traditional or universal humanistic social values.

2.2 Comprehensive results discussion

We initiated two rounds of citizen participation discussions, collected extensive opinions and held specific policy debates to address the impacts of AI on society. Citizen participation can serve as a model for public-private collaboration. Individuals and stakeholder groups closely related to the issues participating in the discussions represent the people and reflect professional opinions from all segments of society. They are sufficient to serve as the legitimate foundation and primary reference for national governance. The following points are supplemented with illustrations：

(I) Individual level: labor force release and empowerment

Individuals believe AI will significantly impact working conditions and rhythm, regardless of occupation, industry category or professional field. It is necessary to consider whether the dividends of labor force release are returned to workers or employers. On an individual level, it is essential to continuously train individuals to acquire basic operational and information recognition skills. This will enable them to cope with future work demands or life transformations. Active empowerment relies on individuals’ professional capabilities within the community, as well as educational training and a variety of citizen participation opportunities. In this manner, the people’s abilities can potentially improve and contribute to social needs.

(II) Stakeholder community level: industrial upgrading and transformation counseling

For the stakeholder communities closely related to this issue, AI is believed to upgrade core products, such as the information and communication technology (ICT) industry, while promoting industrial and social transformation. Some industries, such as design, translation and graphics, rely on counseling to improve organizational capabilities. Most participants mentioned that law, education, imaging and other fields may be significantly impacted. Under the AI wave, the competitiveness of industries will depend on whether workers entering the labor force possess the necessary digital skills, or if there are training mechanisms at the national and community level. Participants also proposed and affirmed the importance of increased collaboration between the public and private sectors, and the use of community support to provide oversight as well as checks and balances.

(III) National level: institutionalization of regulatory policies

Most people believe regulations must be clear at the national governance level so that innovation, research and development can advance in accordance with the law. That is, state regulation must shape the system so that those subject to regulation or related industries and the people can understand boundaries and have the right to express opinions on regulatory boundary formation. Therefore, if the national governance tool design is more transparent, it will benefit individuals, stakeholder communities and even the overall positive cycle of society.

2.3 Potential cooperation methods

The two Democratizing AI Futures Deliberative Workshops were initiated from the concept of Ideathon to launch the Democratizing AI Futures dialogue. As a result, most participants in this event are the target audience for Ideathon activities. The discussion content has gradually shifted away from the impact of AI on various aspects of society and toward policy tools, such as public governance, citizen participation, and data collection methods. Through the review and discussion design, as well as the guidance and leadership of the facilitators, we now have a better understanding of how the social impact of AI is being fermented in Taiwan society.

Please refer to the list of invited participants for future suggestions. Those with more practical insights or implementation experience on the social impact of AI can continue to follow up or arrange in-depth interviews regularly. These key figures may be able to raise awareness for the issues in respective fields, allowing the moda to connect the dots and continue to collect implementation status data from all walks of life. Focus discussions or workshops geared toward citizen participation can be held for specific occupations or industry categories as needed to strengthen the link between public and private sector cooperation, which enables sustainable implementation of Democratizing AI Futures in areas of need from policy planning down to the social demand response level.

2.4 Results report and related links

1. Polis
2. Talk to the City
3. Talk to the City report

3. Summary

3.1 Process overview and results summary

Introduction： In July 2023, the Democratizing AI Futures dialogue was launched using the Ideathon platform, and the Democratizing AI Futures Deliberative Workshop sessions were held in Taipei and Tainan Cities in August and September of the same year, respectively. The event brought together AI experts, scholars and the people to discuss the societal impact of generative AI and explore solutions through continued dialogues.

3.2 Execution team background review

Center for Innovative Democracy and Sustainability (CIDS), National Chengchi University (NCCU) was commissioned by the moda's Department of Digital Strategy to hold two Democratizing AI Futures Deliberative Workshops. The Center comprises team members from the public sector, practice fields and academic circles dedicated to improving democratic participation procedures and social dialogue methods. The center's vision for social dialogue is to start with the issues and progress through various stages of communication and collaboration with social partners. Specifically, social partners with policy interests, such as citizen activists, organizational groups, experts and scholars, and government departments, engage in cross-field or cross-level dialogue via discussions, consultations and joint actions, to confront today’s complex issues, build consensus and promote democratic participation.

CIDS members' creative thinking increases public participation channels in public policy. The center's recent research has focused on the governance of technological subjects, such as NIMBY technology facilities and human rights, social impact assessment on technological innovation, nuclear waste site selection and citizen participation. The goal is to bring public issues to the table to find the best solution for social development after integrating opinions from various forums.

3.3 Implementation motivation and poblem awareness

Generative AI combines Big Data computing, real-time content production, and machine deep-learning technology to develop the ability to imitate and create. The application of AIGC has the potential to revolutionize modern society’s labor structure, industrial layout and even daily habits, bringing significant impacts and opportunities to social operations. According to GitHub, the world’s largest program code management platform, 92 percent of American program developers indicated they use generative AI to write programs at work, and 70 percent of interviewees believed that using AI can improve programming quality and work efficiency. In the past, however, most people in the art and creative industries regarded human intellectual property worthy of legal protection. Nowadays, combining generative AI with personal creativity can vastly increase efficiency and precision, but may also trigger debates on intellectual property ownership. Moreover, regarding the negative impact on society, if people’s social biases and prejudices can be generated using generative AI and quickly spread through the communication system, discrimination will be accelerated to some extent, and ideological dissemination and social antagonism may escalate. These factors must be carefully considered as we develop and utilize generative AI.

Generative AI is rapidly evolving and becoming a competitive field among major technology industries. The development of generative AI has marked the beginning of a new era in society, and the international community has also begun to consider and propose related control plans. For example, the Organization for Economic Cooperation and Development (OECD) defines AI as "machine-based systems that can predict and provide suggestions or decisions to affect the physical or virtual environment to achieve human-set goals. This system has a certain degree of autonomy and is a machine capable of performing human-like cognitive functions." It also proposed the “OECD AI Principles,” which include recommendations for national policies and international collaboration. These recommendations are as follows：(1) Invest in AI R&D; (2) Build an AI-friendly digital ecosystem; (3) Create an AI-friendly policy environment; (4) Cultivate human resources and prepare for labor market transformation; and (5) Implement international cooperation on trustworthy AI (OECD, 2023 17). The OECD research report also said that generative AI can filter large amounts of data and assist governments in making policy decisions. It can predict and help prevent election interference events and improve public services. The OECD agrees that generative AI should be used in government governance, but cautions that it may undermine democratic values by perpetuating and amplifying social inequality. Generative AI may further undermine trust and social contracts if not adequately regulated.

In the context of Taiwan, the impact of generative AI on society is evident as per the international trend, and AI’s interactions with democracy, society and politics are also apparent. However, most current discussions in Taiwan continue to center on industrial development, such as evaluating the use of AI by the private sector to cut costs or create benefits for enterprises. Beyond the development and economic aspects, the impact of generative AI on democracy, society and politics should be critical questions for governance and long-term social development. Regarding academic research development, many studies have recently raised concerns about the various ethical threats that generative AI may pose to the public sector. Since different technological waves will result in other aspects of government transformation, they will also generate other public services and values. On the other hand, research directions that are policy-responsive, social demand-oriented, and focus on social entities or different industrial categories have resulted in more policy-responsive and socially demand-oriented research directions.

Accordingly, NCCU was commissioned by the moda's Department of Digital Strategy to collect data from netizens who are interested in discussing AI issues using the ministry's polis.tw public opinion survey report. The goal is to evaluate the more divergent ideas about specific generative AI issues among the people. Taiwan-specific issues can be addressed through such discussions, enabling policymakers to identify and mitigate potential risks in the future.

3.4 Limitations

(I) Definition of participant representativeness

The Ideathon platform initiated the two Democratizing AI Futures Deliberative Workshop sessions and launched the Democratizing AI Futures dialogue concept. Most participants in this event have contributed to Ideathon activities and volunteered to attend the workshop. The remaining participants are experts and scholars who were invited to attend. The source of participants in this event is specific target groups solicited via the voluntary and invitation systems. Not all event participants are front-line users. It is also clear that the proportion of people working in the design field is significantly higher than that of other fields, and there is a gap in their understanding, ideas and performance in technology applications. The appropriate source, number and representativeness of participant invitations depend on the purpose of the event design because one of the topics to be discussed in these two sessions is revised and used as the submission method for the next Ideathon. The composition of the participants in this event is connected and aligned with the purpose of the discussion. The event’s design attempts to set the agenda by focusing on the impacts of AI on society. Various professionals can provide perspectives to create a preliminary blueprint in this direction. If there are plans to hold future discussions on similar topics, it would be helpful to consider making more diverse designs that consider the participants’ representativeness. This could involve discussions in separate groups, with an invitation-only system for some sessions and a voluntary participation system for others. Such an approach can assist the event organizer in achieving the event’s purpose and reaching the necessary opinion sources.

(II) Agenda setting for deliberative workshops

The two Democratizing AI Futures Deliberative Workshop sessions started by collecting opinions from netizens. These opinions were summarized using polis.tw and supplemented with domestic and foreign news, research trends and other relevant information. The opinions were then categorized into six major themes (internal government governance applications, emerging law-supervision mechanism rule, information identification, data openness, intellectual property, education and human cognitive impacts) for on-site discussion. As the majority of participants in the Ideathon are volunteers, with some invited experts and scholars, the executive team assigns experts and scholars to groups based on discussion topics that align with professional fields. Since the voluntary contributors came from various backgrounds, they were randomly assigned to the topic groups. As a result, the opinions expressed in each group range from academic and practical to imaginative; the participants may not have a complete grasp of the assigned discussion topics. The practical operation of AI is limited by the technology’s immaturity and lack of understanding among most people. As a result, many lack a comprehensive understanding or sufficient experience in dealing with the limitations and dilemmas of practical AI operations. The issues raised by AI applications are not about the knowledge or ability that can be manipulated through citizen consensus. It takes people with actual contact, training, or application experience to conduct a more in-depth analysis and propose prospective recommendations. If the opportunity to discuss similar topics arises in the future, we can prioritize the discussion topics, choose a singular topic or correspond to a specific occupation or industry category and conduct a more in-depth discussion. Only then can we gain a more accurate grasp of AI’s practical applications or impacts on society.

(III) Policy recommendation tracking and implementation

Since citizen participation activities require significant pre-launch preparation and on-site communication, participants who are willing to spend time participating will likely have a certain degree of participation efficacy. There is a sense of efficacy or a certain expectation that the insight offered through discussions will be adopted. Even if the deliberation team has prior knowledge and on-site leadership, a good deliberation event can focus everyone’s opinions on several meeting conclusions that most participants present are interested in and agree with. It may also be able to produce relevant policies through guidance. Even if the discussion during a meeting is engaging and effective in participation, the implementation of actual policy agendas and the impact of meeting conclusions on policies must still be executed by a civil service system with the appropriate powers and responsibilities. Common sense suggests that it is difficult for participants to continue tracking the subsequent policy implementation effects after similar discussion activities. Most policy development and promotion take time and are unlikely to be immediately enforceable, which has policy-oriented or social implications. The complex nature of such development issues makes it harder to confirm subsequent effects. Whether the outcomes of discussions held in event venues will have a tangible impact on policies, or if the content of the discussion aligns with the intended purpose is yet to be determined. Ultimately, this can only be tested through practical implementation. The results of these docking policies and practices are beyond the scope of what can be determined or guaranteed by current deliberation activities.

4. Methods

4.1 Research methods and member participation analysis

The two events drew a total of 95 participants. Regarding gender ratio, males had a slightly higher proportion than females, with the former accounting for 61.1 percent and latter 38.9 percent. Based on two independent analysis sessions conducted in Taipei and Tainan Cities, the male ratio remains higher than the female. Please refer to the figure below：

In terms of age distribution analysis, when the number of participants in both sessions is combined, the majority of participants are between the ages of 18 and 29 (about 40 percent), followed by those between the ages of 30-39 (24.7 percent), and those between the ages of 40 and 49 (22.4 percent). Participants under the age of 49 account for over 87 percent of the total. Most of the participants in this event are from the youth and young adult generations. Please refer to the figure below:

When we combine the number of participants from both sessions and analyze the types of units they belong to, we find that 41.9 percent of the participants are from schools or government think tanks, while 40.9 percent are from civil society groups or companies. The remainder belong to government units. Participants in this event are evenly divided between academic units, as well as the second and third sector. The first sector accounts for 17.2 percent of the total. Please refer to the figure below：

When the number of participants in both sessions is combined, an analysis of the participants’ professional fields distribution indicated that over 60 percent come from the fields of digital technology and design, with the remainder coming from the social sciences, law, business, medicine/psychology and education fields. The professional fields are heavily concentrated in digital technology and design, likely related to the fact that this participant’s invitation source is primarily associated with Ideathon contributors. Please refer to the figure below：

4.2 Process details (topic framework and discussion design)

CIDS and NCCU advocate a social dialogue method that begins with analyzing the issues, inventorying the issue’s current status on the network and identifying the challenges or problems that must be addressed. We then enter the stakeholder identification phase to consider who the stakeholders are and the purpose of each stakeholder’s role in the context of this issue. Next, we can develop and customize operational methods based on practical needs during the issue diffusion phase through various levels of dialogue frameworks. The center’s social dialogue design maintains deliberation and speculation, as well as emphasizes the importance of empathy and putting oneself in the shoes of others. We believe it is vital for every discussion to focus on theory and incorporate physical sensations in relating the issues to everyday life.

This workshop adopted the center’s established methods to inventory and analyze the current state of generative AI issues. Additionally, opinions collected from polis.tw were examined to identify what the people think about the generative AI scenario. Finally, we investigated the causes of debates on various issues through the design and operation of the two deliberative workshop sessions, which can be used as a reference for future policy planning. The following is a brief description of the design concept and operation plan for this event：

(I) Scenario setting

We employed action drama performances at the start of the sessions to help participants easily understand the objectives of the subsequent topic discussion. The scenario setting of the action drama is as follows：

How will the multiverse of generative AI change our current lifestyle?

Scenario – a certain company

"A" owns a consulting company that assists the private sector in developing generative AI services. Generative AI has advanced on a large scale over the years. In addition, generative AI development requires high professionalism. With abundant resources, "A’s" capital has also increased significantly in the past few years. Recently, "A" helped company "B" to develop an image design robot. Employee "C" of company "B" is thrilled that generative AI can help to generate ideas, which has dramatically improved his work efficiency and eliminated the dark period of the past when he had to work overtime.

Employee "D" stumbled upon a photo competition advertisement while wandering around the company after lunch. He was eager to participate because of the attractive prize, but the deadline was less than 10 hours. He had no time to create his own images due to his busy work schedule. So, he decided to use generative AI to help generate and modify the images. Sure enough, it took employee "D" less than 6 hours to revise and submit the images.

Unexpectedly, reviewer "E" recognized Employee "D’s" work as co-generated using generative AI. Reviewer "E" further discovered that part of the content generated by employee "D" came from the works of others, but failed to clearly indicate the source, which may violate the publishing work ethics. So, reviewer E objected and requested the organizer to disqualify employee D.

Scenario - school

In the multiverse of junior high schools, teachers use generative AI as teaching materials and tools, and students are also accustomed to using generative AI to learn and complete homework. For example, student "F" stated that whenever he had a question that he could not figure out, he would open ChatGPT for an answer. Teacher "G's" challenge was how to keep students from overusing generative AI – a common issue for all teachers in the multiverse.

One day, student "F" announced in class he had written a code using generative AI that could change the teacher’s photo to look like Lin Chi-ling, a Taiwan actress and model. Student "F" was very proud of developing this easy-to-use software and shared it with his classmates. One day, classmate "H" got bored and began to modify images using this software. He created a variety of materials for fun. One was to use classmate "F’s" portrait to create a crying image. Classmates passed it around and unexpectedly reached the mobile phone of classmate F’s Parent I, causing a commotion.

Parent "I" even complained to the school’s academic affairs office, assuming "F" was bullied by his classmates. Parent "I" walked into the academic affairs office and demanded dean "J" rectify the situation. At this time, dean "J" was using generative AI to create office documents. The multiverse’s public service system has already created a generative adversarial network capable of checking whether application information is correct and completing the tasks quickly with technical assistance. After hearing the entire story, dean "J" realized that the students had gone too far with generative AI. Dean "J" began to consider ways to prevent students’ perceptions from deviating from reality and whether it was time to revamp the teaching model.

(II) Table assigning and topic setting

After the action drama, the participants were divided into six tables, each covering one of the following topics：governance (internal applications), emerging rule of law (supervisory mechanism), information identification, data openness, intellectual property and education and human cognitive impact. The Q&A process at each table first defined each topic’s challenges and issues. Participants first discussed and identified the benefits, challenges and risks associated with increased affinity and trust in generative AI applications, and then explored the possible solutions. The issues to be discussed at each table are first explained as follows.

(III) Discussion model and tool sheet design

For this session, we designed the discussion process based on the preceding topic framework and discussion concept, as well as incorporated elements such as topic knowledge, empathy and emotions. Given that the majority of the participants are AI experts or members of relevant communities, most have professional research or empirical insights into the issues. Compared to the people's discussion model, we expect more participants for this session, and the discussions will be more focused and actionable.

We focused the discussion model on specific audiences and maintained an open discussion style. Instead of an open discussion with no limits on the topic or argument, groups should conduct case studies on various topics. After the case studies, the table leader (facilitator) organized the members’ views to ensure that everyone understood the topic and the discussion process could continue smoothly. Participant discussions were used to collect valuable opinions and perspectives on the topics.

The composition of participants in each group is roughly summarized in the figure below. It includes the polis.tw community; information tool users, experts and scholars in related fields, as well as dialogue based on public governance perspectives. This discussion aims to envision new public governance methods and reach a consensus. To achieve this goal, we emphasized the importance of having a public governance perspective in addition to the participants’ perspectives, like experts or community members. This approach can help achieve the desired discussion outcomes and enable the application of results effectively. The design for this part was also used and implemented for the discussion tool sheet and facilitator leadership skill guidance.

Participants in this discussion were divided into six groups/tables based on the topic. Ten people were assigned to each table, including six representatives from relevant communities, two experts and scholars, and two facilitators to lead the discussion flow and guide the intervention from a public governance perspective. There were 60 viewpoint contributors in a single session, as shown in the figure below：

The structure of this group discussion is roughly divided into two major parts. The first round “inventoried the current situation, advantages, and challenges of AI’s impact on the society” while the second round focused on “how to use policy tools to promote coexistence between AI and humans.” At the deliberation and design level, the first round mainly involves presenting questions and inviting the participants to consider the relationship between themselves and the topic. It is also the warm-up for the topic discussion process, which creates opportunities for mutual understanding and prepares each group to discuss and analyze the assigned topics further. Group sharing is held at the end of each discussion round to allow participants to understand cross-group opinions and construct a more complete imagination of the topic.

During the second round, the facilitator guided the formation of the most basic classification based on public policies according to the views shared across the groups in the morning. The morning session addressed how policy tools can respond to risks and challenges by adopting a top-down public governance and bottom-up citizen participation model. After analyzing and sequencing the conditions, we formulated action plans that are both executable and innovative to usher in a new era of coexistence between humans and AI.

Based on the session’s purpose and discussion plan, our team created a tool sheet to help advance the on-site discussion. Participants are asked to share their positive and negative thoughts on AI applications using indicators. The participants of each group can observe each other and understand AI’s preferences and application boundaries by reviewing the records and summaries on the record sheet.

Then, we entered the topic discussion phase primarily associated with AI’s impact on individual aspects, including advantages and challenges. We used the social structure classification framework to guide participants in sharing perspectives on personal cognition, stakeholder communities, social influence and national governance. The goal is to estimate the different needs and challenges faced by agencies at various levels of society. During this phase, we went beyond the question-asking level by guiding the participants to expand their original thinking framework and imaginations about the issues.

The ultimate goal for the second round of discussions is to propose possible co-creation solutions. After discussing the topic in the morning, each group shared their understanding and the various problems and challenges that must be addressed. Before proposing solutions, we reset our expectations to re-energize the discussion. Although we all know there are many challenges and problems to solve, it is important to consider why humans still aspire to build a future society where AI coexists. What policy tools are currently available to help achieve this ideal? At this level, we categorized policy tools into three major categories：laws/rules, strategies/plans and competition projects. The goal is to assist participants in organizing solutions into specific categories, allowing them to effectively link to the corresponding policy development stage during policy implementation.

Prior to implementation, our team usually conducts internal drills via the facilitator’s workshop to establish the design framework and identify issues that can meet the original design vision while meeting the operation units’ expectations. Drills also ensure facilitators can implement the framework and connect the issues and practices. The facilitators will guide the participants step by step to share opinions during the on-site discussion. Table records typically include writing keywords from the participants’ opinions on N stickers or colored pen drawings, allowing participants to look back and construct their own views through opinion stacking. Finally, the facilitator summarized and organized the opinions to form specific suggestions within the group.

4.3 Online discussion process input and case examples

Polis analysis

Before the event’s design deadline, Aug. 1, 2023, 403 netizens from the polis.tw poll voted and made 8,591 choices. On average, each voter made over 21 choices and proposed 126 arguments, and a total of 260 people were divided into groups. The following briefly describes the opinions and preferences for generative AI applications expressed in the polis.tw poll：

(I) Analysis of opinions from those with a high degree of consensus

In general, netizens’ consensus-oriented statements fall into two categories. The first represents a more optimistic view of AI services. They believe AI applications are just tools such as auxiliary or collaborative products, and there is no reason to be concerned. They agree that AI can help humans work more efficiently, but the benefits and drawbacks vary depending on how humans use it. The opposing viewpoint believes that AI has the potential to alter certain social conditions and raises doubts about certain regulatory boundaries.

No.	Content	Degree of Agreement
118	The referenced sources of AI-generated content should be appropriately labeled.	95 percent agree (N=145)
44	As AI technology advances toward new boundaries, the lines between human and machine-generated creations become increasingly blurred. Key intellectual property rights and legal liability questions must be explored in depth.	94 percent agree (N=155)
54	Public servants must receive relevant training before using AI to handle their duties.	92 percent agree (N=166)
104	We cannot rely solely on AI. Regarding security protection and judgment, human nature must also be considered. This is the biggest difference between humans and AI.	92 percent agree (N=157)
45	The current development status of AI will dramatically alter the ecology of many industries.	91 percent agree (N=155)
48	Regarding AI standards, different cultures and contexts worldwide must be considered rather than just the views of a single country.	88 percent agree (N=171)
55	The use of AI can reduce the burden on national civil servants, but it is critical to consider how to handle confidential information without risk of leakage.	87 percent agree (N=165)

(II) Analysis of opinions from those with a certain degree of consensus

The three groups from this report indicated that Group B has relatively neutral opinions, group C is more optimistic about AI tools, and group A pays more attention to human interaction and believes that AI will partially change the current status. After analyzing the opinions of those with a high degree of consensus, we referred to the opinions that all group members agreed with cross-group consensus to confirm the opinions with a high degree of consensus and used them as an important reference for subsequent discussion topic selection. The current observation is based on the degree of consensus, which ranges from high to low but does not deviate from the framework of the preceding analysis. One phenomenon we noticed is that most people who are pleased with the results and those who believe the current situation will be partially changed are not mutually exclusive. Some opinions have also reached cross-group consensus. The opinions with group consensus are listed in the following table：

No.	Content	Degree of Agreement
44	As AI technology advances toward new boundaries, the lines between human and machine-generated creations become increasingly blurred. Key intellectual property rights and legal liability questions must be explored in depth.	94 percent agree (N=155)
80	Use AI with a positive attitude and learn how to use it effectively together.	94 percent agree (N=169)
85	The application of AI is still in its early phase. There is still much room for deciding whether AI poses an opportunity or a threat. Limiting yourself is unnecessary, but you must prepare for AI in advance.	95 percent agree (N=185)
117	AI is a tool, but ensuring the authenticity of the information provided is the expression of personal professionalism. People who use AI will improve their work efficiency. Similar to the rise of Google, the rise of AI is another change in the thinking model.	100 percent agree (N=27)
45	The current development status of AI will dramatically alter the ecology of many industries.	91 percent agree (N=155)
104	We cannot rely solely on AI. Regarding security protection and judgment, human nature must also be considered. This is the biggest difference between humans and AI.	89 percent agree (N=77)
83	The final judgment regarding AI matters still belongs to humans.	92 percent agree (N=152)
54	Public servants must receive relevant training before using AI to handle their duties.	92% agree (N=155)
106	AI cannot replace all life skills but it can make life more convenient.	94 percent agree (N=37)

(III) Aspects of more divergent opinions

In response to the questions with highly divergent answers expressed by the opinion survey participants, such questions may be discussed in the current social context. They also serve as an important reference point for analyzing future planning discussion topics. The questions with divergent answers are listed as follows：

No.	Content	Degree of Agreement
56	It is trustworthy for the government to use AI to carry out its duties.	36 percent agree; 5 percent disagree (N=48)
47	When AI becomes ubiquitous in the future, the entire world will transition from the post-truth era to the no-truth era, altogether abandoning or unable to pursue the truth.	42 percent agree; 36 percent disagree (N=183)
10	The fear of new technology is exaggerated, and humans do not need to understand how it works to take full advantage of it.	55 percent agree; 34 percent disagree (N=185)
66	Overreliance on AI may deprive students of interpersonal communication opportunities and compromise their social skills.	53 percent agree; 25 percent disagree (N=189)
60	AI destroys and compresses the original data, making it more difficult to obtain accurate data during the learning process.	45 percent agree; 27 percent disagree (N=156)
39	When AI creates something that humans used to make, it becomes more meaningful.	60 percent agree; 20 percent disagree (N=163)
69	I believe the works submitted can be generated by AI and then completed using Photoshop.	66 percent agree; 15 percent disagree (N=178)
6	Copyright costs will become less constrained as the production of large-scale, high-quality derivative works becomes more affordable.	64 percent agree; 22 percent disagree (N=166)
67	I believe there should be a separate award for works generated using AI	66 percent agree; 15 percent disagree (N=48)
64	As educators, we must provide experiences that are more authentic and personal than AI-generated content.	63 percent agree; 13 percent disagree (N=165)

(IV) Suggested discussion topics

According to the preliminary analysis above, the netizens’ attitudes toward generative AI can be roughly divided into two major groups based on majority votes. One group is happy with the current results, and the other believes some control is required. When the consensus across the groups is further classified, there is no significant conflict between opinions, and the majority consensus is generally agreed upon across the groups. According to the survey participants, we can tentatively regard a “positive view towards generative AI, but it is necessary to propose control plans for certain industries” as the majority opinion from this survey.

As we delve deeper into the divergent parts, we can identify the issues that cause confusion or doubts among users, such as government governance, including internal applications and regulatory mechanisms, information identification, data openness, intellectual property, education, and human cognitive impacts. These concerns are likely raised by those who support generative AI but believe some regulations are necessary. Since the same group clearly disagreed on choices for the topics listed above, the cause should be investigated further using qualitative data collection.

Based on international trends, the current state of domestic issues and opinions gathered from polis.tw, we recommended the topic discussion workshop to focus on six topics for the discussion topic design. They include government governance, focusing on internal applications; the emerging rule of law, focusing on regulatory mechanisms; information identification; data openness; intellectual property and education; and impacts on human cognition.

Following the preceding analysis, professional facilitators distributed the relevant issues to the people for on-site deliberative workshop discussions.

4.4 Methodology design and implementation

In a democratic society, representatives are elected by ballots and regular elections using direct or representative democracy mechanisms to exercise public power on behalf of the people. Although the vote aggregation and competitive democratic operation model can achieve the operational effectiveness and decision-making efficiency required for national governance, it is not without flaws. When a majority decision-making system and winner-takes-all approach are in place, minority citizens may struggle to express opinions fully, resulting in what is known as majority violence. To address the challenges of aggregative democracy, contemporary scholars have proposed deliberative democracy, which advocates allowing citizens to participate in public policy discussions and form collective opinions to influence decision-making through public deliberation.

With the rapid development of contemporary society, public issues are becoming more complex and cross-disciplinary. As a result, the people's demand for information disclosure is increasing. Policy planning and promotion require more than just seeking compromise or consensus between the government and society. Instead, it is becoming increasingly important to address the diverse differences between social groups. In today’s world, there is a pressing need for a platform that facilitates communication and interaction. This is necessary to encourage more discussions and diverse viewpoints, and to establish new collective opinions or consensus through mutual understanding and observation. Such a concept coincides with the so-called deliberative democracy. Effective social dialogues are crucial for enhancing democratic understanding and improving public discourse.

The design of these two event sessions is titled Deliberative Workshop because it heavily references the spirit of deliberative democracy. We invite citizens to participate in agenda discussions, express opinions, and organize and summarize the suggestions discussed in the public policy planning and design process, which the citizen participation model supplements. Of course, such a time-consuming process cannot be automated solely through meetings due to time constraints. Instead, professional facilitators must lead the discussions to help participants quickly understand public issues and reach conclusions within a limited time frame. By setting sufficiently specific questions, we can achieve our goal of gathering constructive or implementable policy recommendations.

As previously stated, the executive team of CIDS and NCCU has experience discussing numerous issues during deliberation operations and emphasizes the use of perceptual empathy. When promoting deliberative democracy, we believe it is critical to emphasize mutual understanding and openness by listening to the opinions of others to reflect on and broaden the people’s perspectives. It is crucial to highlight such experiences as we promote deliberative democracy. The deliberative workshop also included play design elements in which team members performed short plays about the impact of AI on society, allowing participants to imagine possible scenarios and relate them to their own experiences. This enabled subsequent discussions to transcend the theoretical level and be closer to human society’s actual operation patterns and needs.

4.5 Methodology impacts and side effects

As mentioned above, these two workshop sessions’ activity design and on-site operations closely followed the spirit of deliberative democracy. The goal is to replicate the benefits emphasized by the deliberative democracy model. The implementation team conducted research and presented the findings in a simplified and easy-to-read format. We then established the agenda and discussion topics so that participants could speak from multiple perspectives in a limited time. The opinions were then collected and analyzed under the facilitators’ leadership. We considered the topic’s perceptual inspiration and used play performances to broaden the participants’ perceptions. This elevated the discussion to more than just words on paper, making it more relevant to potential situations in human society.

This method of operation has clear advantages. It allows all participants to express opinions within a specified time frame under the guidance of skilled facilitators. Additionally, it enables participants to communicate, reconsider subjective views, and strive for a collective consensus by observing one another. Setting sufficiently specific and clear questions created opportunities to collect constructive and actionable policy recommendations. By actively participating in discussions that foster mutuality, reflection and acceptance, the participants can develop empathy and an accepting attitude toward the on-site recommendations. The positive cycle achieved through deliberation can assist public decision-makers in promoting policies, regardless of whether the issues discussed on-site can return to the decision-making level or are in the policy planning, legalization, or implementation phase.

In contrast, the side effects can be roughly divided into two aspects. The costs of social mobilization for activities such as deliberation and discussion are considerable. For example, only a few people have the time to participate in the entire process. If the quality of the discussion is to be considered, inviting an excessive number of participants to a single event is inappropriate. The usual size of an event is about 20-50 people. Obtaining comprehensive opinions without holding multiple events may be challenging due to the limited number of people who can attend a single event. If a deliberative democracy event has specific social representation requirements, systematic or purposive sampling may be used to invite participants to mitigate the risk of participant composition influencing the opinion collection.

The second side effect is that the meeting’s conclusions must be verified. Because of the high cost of managing deliberative democratic activities, both the organizer and the participants typically have high expectations for the implementation and impact of the conclusions. If the discussion concerns policy issues, the relevant conclusions must still be returned to the appropriate administrative agencies for implementation. Only then can we test whether the results discussed on-site are feasible for implementation. Retesting the method discussed may not be possible until the time of implementation. This is most likely because in the deliberative democracy discussion process, no matter how positive the atmosphere is or how remarkable the results are, it is difficult to claim that the meeting’s conclusion reached by the collective consensus at the time of the meeting will be effectively implemented during policy design.

To mitigate the abovementioned side effects and risks, when the organizers realize that the topic expected to be discussed is not a general social issue but a contentious issue that may soon become a policy agenda, it is necessary to ensure that the invited participants can express perspectives on the agenda. Meanwhile, facilitators must serve as intermediaries, sort through and summarize the topic materials, and propose agendas with discussion value. The facilitators must also navigate diverse opinions on site, understand the meeting’s purpose and develop specific and constructive solutions to facilitate public discussions on the value of deliberative democracy effectively. It is also essential to connect these discussions to policy practices and exert influence to ensure they have a meaningful impact.