---
tags : vTaiwan小松, CAI, ruleAI, OpenAI
---
# {20230608} vTaiwan 延伸討論
時間：{2023/06/08} 週三 21:00-23:00
地點：
出席：
線上組：
線上參與連結：https://meet.jit.si/vTaiwan

Todo
---
FtO jeju CHAT G0vParticipateTaiwan
https://docs.google.com/presentation/d/1kqGM4qXB5xZL5XYuxITF4e9lS--5RBSrF1idWdAkbiM/edit?usp=sharing

### [Application Questions/Draft Application](https://hackmd.io/4qKOpMJ1SiK9Y6FuFodTVQ)

I have tried to capture the below in the application document, and copied the questions the bid proposal must include. To discuss soon:

1. Group Discussion: Project Partners - when?
1. Shape of the practical experiment in applying AI to the process.
1. Evaluation/comments on [first draft of Application](https://hackmd.io/4qKOpMJ1SiK9Y6FuFodTVQ) (above)
1. Resourcing plan - how to spend the grant funds

合作對象
---
Partnership 會是一個蠻重要的評斷條件，vTaiwan 的力道是可以很方便的找到合作。

Carl Miller：可能的 connection來源
Alex Krasodomski, Chatham House

Moda RR(id: delightfulchaotic) Mashbean(黃豆泥)

eg. 德國組織 合作
Digitalisierung und Gemeinwohl
Dr. Felix Seiker

CHAT GPT
G for g0v
P for Participant
T for Taiwan

Chat!"G0vParticipateTaiwan"

分享
---

[好讀版](https://www.craft.do/s/HiXip841HxtfxN)
內容貼過來：

### vTaiwan process design

#### Introduction to vTaiwan

vTaiwan is a civic tech initiative that originated in Taiwan. It is an online platform that engages citizens in policy discussions and decision-making processes.

vTaiwan is an initiative that originated in Taiwan’s civic tech community. It provides an organic (flexible, iteratable) and hybrid process (in-person and virtual engagment) that engages citizens in policy making process, espeically in digital regulatory reform. It proves particularly valuable in addressing emerging technologie, inviting the public, and topic-relevant stakeholders in discussions regarding the necessary regulations.

Notable instances where vTaiwan has been instrumental include the regulation of UberX and Unmanned Aviation Vehicles (UAV) where the process facilitates diverse inputs from large participation and delivers bridging-based statements that informs principles in policy making.

#### Discussion around AI

As AI continues to advance at a rapid pacey, it's becoming more and more important to have meaningful discussions about the rules and regulations that should govern its use. Without doing so, we risk having extreme measures such as outright bans or unbridled deployment without adequate safeguards.

NAIAC Year 1 Report and NIST AI Risk Management Framework both emphasize the inclusive discussion from different communities, especially unrepresented communities in different steps of AI life circle. We believe vTaiwan, with the integration of civic discussion

This is exaclty where vTaiwan comes in. It recognises this familiar tension and the process has been specifically designed to effectively navigate and address controversal issues that arise with the emergence of new technologies.

By leveraging the vTaiwan process, contentious topics related to AI can be examined and discussed, allowing for a comprehensive exploration of various perspectives and potential solutions.

I propse us from the vTaiwan community consider this work in two excitig directions:

1. How can the vTaiwan process facilitate discussion on the regulation of AI?

This involves exploring how the vTaiwan process can be adapted, refined, or reshaped to effectively address the specific challenges surrounding the regulation of AI. Given that:

a. AI could be a broad topic refering to multiple aspects, such as refering to its algorithms and training methods, the interaction with AI-powered interfaces, and identify and personhood of AI etc

a.A1 - create an interface of AI to explain what AI is about, using tools like chatbots to create a beginner-friendly environment, as well as lowering down the pressure for other participants.

b. Unlike cases like UberX where stakeholder boundaries were relatively clear, AI has the potential to impact various aspects of our lives, necessitating the involvement of a wide range of stakeholders and their input.

b.A1 - currate an open list of different aspects of live (food, education, entertainmanet, busienss, transportation / minority groups) AI tools can help summarize the

c. The risk and challenges AI brings can be related to multiple different agencies in the government. How to help government navigate this interdiscplinary challenge to regulate and make AI trustworthy is another direction we need to consider when adjusting vTaiwan process. i.e. The advance use and deployment of AI in medical research may be a issue concerning Ministry of Health and Welfare, National Science and Technology Council, and personal data related agency.

c.A1 - Exploring the potential and possibility of vTaiwan mechanism to include more than one governmental agencies to participate the discussion. Using voice recognition, script transmition and summarization tools to help accelerate and elevate the quality of discussion between different agencies, as well as lowering down the workload for the civil servants to incentivise them to participate into the discussion.

d. For those communities unrepresented, they are also the ones that can not adopt online tools or exploring information necessary for the discussion. How to help those groups overcome the challenges they face when participating the discussion, including the scarcity of time, effort, attention. Moreover, more we do to improve the virtual discussion process with assistance from AI, more disparity appears between us and the communities that haven't catch up the wave of digitalization.

d.A1 - Approaching to the existing tunnels like social workers, local parliament members to get the information of potential representives of these communities.
d.A2 - Visit physically to the existing public infrastructure including local community center, library, or local schools to do the hybrid discussion. The hybrid discussion can become the bridge between the physical world and the virtual world. AI can help recording and translating their voice and give inputs

(more to discuss with the vTaiwan community)

It's crucial to explore these complexities and potential adjustments required for the vTaiwan process.

2. How can AI contribute to addressing existing issues within the vTaiwan process and enhance it for the future (vTaiwan 2.0)?

Consider how AI can play a role in resolving any challenges or limitations currently faced by the vTaiwan process. Explore how AI technologies can be leveraged to improve the efficiency, effectiveness, and inclusivity of the vTaiwan platform. This includes identifying areas where AI can enhance public participation, facilitate information gathering and analysis, and streamline decision-making processes. By exploring the synergy between AI and the vTaiwan process, it is possible to envision an improved version of vTaiwan, that harnesses the power of AI to create an even more robust and impactful platform.

#### Rooms to improve at vTaiwan
**Discussion**

1. How can AI facilitate "cultural translation" to enhance participants' understanding of the topics under debate, thus enabling more informed discussions? (For example, AI could assist creating a map of topic areas and visualse how one has understanding among them, and suggested conversations to dive into, this idea subtly separating learning mode from contributing mode.)
2. How can AI help participants articulate their perspectives more effectively? (For example, an AI-powered conversational app, known as "Critical Friend," could assist individuals in refining their thinking on complex topics through meaningful discussions with an AI facilitator.)
3. How might AI guide participants using the ORID (Objective, Reflective, Interpretive, Decisional) method to overcome emotional elements and focus on constructing well-informed statements and decisions? The application of emotional recognition algorithms is also worth considering in this context.

**Process**

1. How can we better bridge the gap between the virtual and in-person worlds? Considering vTaiwan's aim to integrate online and offline discussions in its third stage, which involves livestreaming in-person meetings to engage online participants and having facilitators read out online comments to in-person attendees.

Ans. AI based automatic tools can figure out the problem in several ways:

(1) AI-based automatic tools like whisper or ChatGPT can generate the script and digest the script from the live stream to summarize different stakeholder's view.
(2) AI-based automatic tools can create real time subtitle for the participants in the physical world, solving the problem that the online users usually have challenges fully understand the discussion when they are watching the livestream or participate in the discussion.

2. How can AI assist in attracting relevant stakeholders to participate in discussions and contribute their insights? Given that every topic can have relevance to different stakeholders, but efforts are required to etablish those links. Can AI help in identifying those links and engaging the attention of stakeholders?

Ans. AI tools can do the search in the academic database to find out the professionals as targets to reach out. AI tools also have the potential to automatically invite the professionals to offer the opinion. The professionals can also use AI tools to give out opinions in a more efficient way.

**Effectiveness**

1. Long-term Tracking: How can AI facilitate tracking the input of each participant, connecting it to a draft bill, and ultimately tracing its impact on the final outcome?

Ans. We can use the system like Hypercerts or other data tracker to track the contribution of each volunteers helping ficilating the discussion,

**Transparency**

1. How can AI contribute to constructing an open and transparent process? Given that tasks like note-taking, note-sharing, and livestreaming have traditionally relied on manual effort, how can AI automate and improve these processes to enhance transparency?

**Commitment**
1. How might we use AI help bridge the vTaiwan process and government commitment? (For example, ensurig trustworth AI that helps the government harvest the benefits and opportunies brought by AI, but meanwhile address problems and challeges brough by AI)

**Basic Principles when we need to adopt AI into civic participation**

＠teemo
* 與真人互動時，並須明確的知道他所使用的人機互動介面正在與非人類溝通。
* 在真人知道人機互動介面正在扮演某個角色時，可以提供情感支持。
* 允許人工智能助手提供減少人類花費冗長重複性時間的服務，或允許人工智能助手提供找到人類疏漏的部分。
* 當人工智能能夠識別推特上的兩個人是同志情侶關係並且交往中時，此類型服務造成少數族群高度焦慮。
* 必須提供目前溫度 (NLP / LLMs中的Temperature) 相關資訊。
* 應該由社群使用民主化機制產生 Moderation / Content Filter / Policy，並且能夠切換。

**Root Problem in vTaiwan**
There are several problems existing in the vTaiwan system right now.

1. There are not any engineers currently maintain the system periodically.
2. There are not any resource for the engineer

Potential solution: cooperate with sch001 to get the grant and realize the project. Then bring back the engineer to maintain it

**Example Questions proposed by OpenAI:**

* How should AI assistants respond to questions about public figure viewpoints? E.g. Should they be neutral? Should they refuse to answer? Should they provide sources of some kind?
* Under what conditions, if any, should AI assistants be allowed to provide medical/financial/legal advice?
* In which cases, if any, should AI assistants offer emotional support to individuals?

* Should joint vision-language models be permitted to identify people's gender, race, emotion, and identity/name from their images? Why or why not?

* When generative models create images for underspecified prompts like 'a CEO', 'a doctor', or 'a nurse', they have the potential to produce either diverse or
homogeneous outputs. How should AI models balance these possibilities? What factors should be prioritized when deciding the depiction of people in such cases?

* What principles should guide AI when handling topics that involve both human rights and local cultural or legal differences, like LGBTQ rights and women’s rights? Should AI responses change based on the location or culture in which it’s used?
* Which categories of content, if any, do you believe creators of AI models should focus on limiting or denying? What criteria should be used to determine these restrictions?