| import gradio as gr | |
| with gr.Blocks() as tutorial: | |
| gr.Markdown("**Practice tutorial: Creating Ethical Charters in AI Development**") | |
| gr.Markdown("This tutorial outlines the process of creating ethical charters for AI development projects, focusing on moral value pluralism and collaborative, inclusive practices. We discuss why it’s important to identify common core values to guide projects and navigate tensions, drawing inspiration from Confucian ethical traditions that promote harmony. The session includes steps for determining relevant values, building consensus, and implementing these values within specific AI initiatives. We aim to promote values-informed AI development by utilizing previous work as case studies, such as the BigScience project and other ethical frameworks from major tech companies. Participants will learn techniques for integrating ethics into their work, ultimately influencing priorities and impacts in AI technology development.") | |
| gr.Markdown("# Description") | |
| gr.Markdown("""The goal in constructing ethical charters is to find common core values that can serve as guidance in a project, operating as a pivot for identifying and articulating different possible priorities and navigating tensions between them. Part of the importance of ethical charters lies in the construction practice itself, as it fosters collaboratively gathering, reflecting, and deliberating together in an inclusive space. | |
| Our approach to constructing ethical charters is based on the ethical concept of moral value pluralism, which allows different value systems present among the collaborators of a project to be considered of equal importance. In line with this approach, we adopt the practice typical of the Confucian ethical tradition that emphasizes, among other things, harmony (和). This guides resolution, focusing on the coexistence of different parties \cite{ConfucianIdeal}. The notion of harmony thus fills one of our primary goals in helping people build ethical charters in AI development: to guide what people ought to do when making choices that affect others. Bottom-up collaboration, the support of ethics as a tool for constructing guidance, and value pluralism based on the notion of harmony make up the beacons of our approach. | |
| Our tutorial will draw on the idea that the scientific domain of artificial intelligence provides an opportunity for researchers from different disciplines to work together towards common goals. Tools of ethical reflection can help guide this interdisciplinary collaboration, aiding in distinguishing opinions from moral judgments, the importance of asking the right research questions, and the complexity of human judgment and emotion when defining "good" and "bad", "right" and "wrong". We will utilize previous work we have led as case studies, for example, the experience of drafting an ethical charter for BigScience \cite{BigScience}, an open science project focused on developing a multilingual language model and its dataset for research purposes. We will also connect to more well-known ethical charters similar to those discussed in this tutorial also include "pillars" and "principles" statements increasingly released by tech organizations, | |
| such as [Google's AI Principles](https://ai.google/responsibility/principles/), which Dr. Mitchell worked on operationalizing Google-internally; Microsoft's [Responsible AI Principles](https://www.microsoft.com/en-us/ai/responsible-ai) and Meta's [Five Pillars of Responsible AI](https://ai.meta.com/responsible-ai/). | |
| ## Impact statement | |
| Participants will learn (1) why ethical charters are important and how they can guide AI development; (2) processes for eliciting values among groups while building consensus and avoiding inconsistency; and (3) techniques for operationalizing these values in projects. This workshop is one contribution in the broader enterprise of building ethical charters in AI development. As such, intended outcomes include the growth of more values-informed development of AI technology, fundamentally influencing what is prioritized and how people are affected. | |
| ## Timeline | |
| We will work through slides presented in-person and virtually if remote participation is made possible, allowing open discussion throughout. | |
| Introduction (\~8 min): What are ethical charters, and why should we build them? | |
| Identifying Values (\~17 min): Participants in a project or an organization do not come to the table without prior moral beliefs. The process of identifying relevant values for a project embraces this, and can involve either (1) mapping previously defined organizational values to a target project, or (2) articulating values based on the beliefs of the people at the table. | |
| Consensus Building (\~10 min): How to build consensus around values and face potential inconsistencies once a set is elicited. | |
| Operationalizing (\~10 min): Techniques for operationalizing these values in projects.""") | |
| tutorial.launch() |