Expert Council on Well-Being and AI

我们组建了“幸福与人工智能专家委员会”，以帮助指导我们持续开展的工作，致力于为所有人打造更有帮助的 ChatGPT 和 Sora 体验（https://openai.com/index/building-more-helpful-chatgpt-experiences-for-everyone/）。这支八人小组汇集了在研究技术如何影响情绪、动机和心理健康方面拥有数十年经验的顶尖研究者与专家。他们的职责是为我们提供建议、提出问题，并帮助定义各年龄段与人工智能互动时应有的健康标准。

今年早些时候，我们就已开始非正式地咨询其中许多专家，比如在开发家长控制功能（opens in a new window）（https://chatgpt.com/parent-resources?openaicom_referred=true&openaicom-did=9b8001fd-7d62-477c-a8bd-efe6eaf11102&openaicom_referred=true）以及当青少年可能处于困境时向家长发送通知的措辞时。随着新委员会的正式成立，我们扩大了搜索范围，纳入了心理学、精神病学和人机交互领域的更多专家，带来了关于人们如何与技术互动以及如何受其影响的新视角。鉴于青少年使用 ChatGPT 的方式与成年人不同，我们也邀请了几位在构建支持青少年健康成长的技术方面有经验的成员。

我们仍然对自己的决策负责，但在构建能支持人们幸福感的先进 AI 系统时，我们会继续向该委员会、全球医师网络、政策制定者等学习。

认识委员会成员

我们感谢这批成员的深厚专业知识与共同承诺，致力于使 AI 更具支持性与安全性。

• David Bickham, Ph.D. — 波士顿儿童医院数字健康实验室研究主管、哈佛医学院助理教授。他的研究关注年轻人社交媒体使用如何影响心理健康与发展。
• Mathilde Cerioli, Ph.D. — everyone.AI（一个帮助人们理解儿童使用 AI 的机遇与风险的非营利组织）首席科学官。她拥有认知神经科学博士学位和心理学硕士学位，研究聚焦 AI 与儿童认知及情感发展的交叉。
• Munmun De Choudhury, Ph.D. — 佐治亚理工学院互动计算 J. Z. Liang 教授。她利用计算方法更好地理解在线技术在塑造和改善心理健康方面的作用。
• Tracy Dennis-Tiwary, Ph.D. — 亨特学院心理学教授，同时是 Arcade Therapeutics 的联合创始人兼首席科学官。她开发用于心理健康的数字游戏，并研究技术与情绪健康之间的互动。
• Sara Johansen, M.D. — 斯坦福大学临床助理教授，斯坦福数字心理健康诊所创始人。她的工作探讨数字平台如何支持心理健康与福祉。
• David Mohr, Ph.D. — 西北大学教授，行为干预技术中心主任。他研究技术如何帮助预防和治疗诸如抑郁与焦虑等常见心理健康问题。
• Andrew K. Przybylski, Ph.D. — 牛津大学人类行为与技术教授。他研究社交媒体和视频游戏如何影响动机与幸福感。
• Robert K. Ross, M.D. — 在健康慈善、公共健康和社区健康倡议方面的全国性领导者与专家，曾任加州基金会（The California Endowment）总裁兼首席执行官，早年职业为儿科医生。

我们将如何与委员会成员合作

上周我们以一次线下会议启动了该委员会，深入探讨 OpenAI 在这些领域的当前工作，并让委员们与他们将要合作和提供建议的团队见面。

我们与委员会的工作将包括定期检查我们的做法，并通过循环会议探讨诸如 AI 在复杂或敏感情形中应如何表现、以及哪些护栏（guardrails）能最好地支持使用 ChatGPT 的人。例如，在推行家长控制功能（https://openai.com/index/introducing-parental-controls/）时，我们咨询了个别成员，帮助我们确定应优先构建哪些控制功能，以及在青少年似乎处于困境时如何最佳地通知家长。他们的反馈影响了我们所用消息的语气，使其对青少年和家庭成员都显得关怀且尊重。

该委员会还将帮助我们思考 ChatGPT 如何对人们的生活产生积极影响并促进他们的幸福感。我们的一些初步讨论聚焦于什么构成幸福感，以及 ChatGPT 如何在各个生活方面帮助人们更好地应对挑战。我们将继续倾听、学习并分享这项工作带来的成果。

扩展我们的安全工作

除了“幸福与人工智能专家委员会”就更广泛的幸福策略提供建议外，我们还在与“全球医师网络”（Global Physician Network）内的一组多学科精神健康临床医师和研究者合作，来塑造我们的模型行为和政策，并测试 ChatGPT 在现实情境中的应对方式（https://openai.com/index/building-more-helpful-chatgpt-experiences-for-everyone/）。这项工作涵盖精神病学、心理学、儿科和危机干预，帮助确保我们的系统基于临床理解和最佳实践。

关于主 ChatGPT 模型的改进以及我们在如何最好地支持人们方面的学习，我们很快会有更多内容分享（opens in a new window）（https://help.openai.com/en/articles/9624314-model-release-notes#h_d7ec15c07a）。

We’ve assembled the Expert Council on Well-Being and AI to help guide our ongoing work to build more helpful ChatGPT and Sora experiences for everyone. The eight-person council brings together leading researchers and experts with decades of experience studying how technology affects our emotions, motivation, and mental health. Their role is to advise us, pose questions and help define what healthy interactions with AI should look like for all ages.

Earlier this year, we began consulting many of these experts informally, such as when we were developing parental controls⁠ and the notification language for parents when a teen may be in distress. As we formalized the new council, we broadened our search to include additional experts in psychology, psychiatry, and human-computer interaction, bringing in new perspectives on how people relate to and are affected by technology. Because teens use ChatGPT differently than adults, we’ve also included several council members with backgrounds in understanding how to build technology that supports healthy youth development.

We remain responsible for the decisions we make, but we’ll continue learning from this council, the Global Physician Network, policymakers, and more, as we build advanced AI systems in ways that support people’s well-being.

Meet the council members 

We’re grateful to this initial group for their deep expertise and shared commitment to making AI supportive and safe. 

• David Bickham, Ph.D.—Research Director at the Digital Wellness Lab at Boston Children’s Hospital and Assistant Professor at Harvard Medical School. His work looks at how young people’s social media use affects their mental health and development.
• Mathilde Cerioli, Ph.D.—Chief Scientific Officer at everyone.AI, a nonprofit helping people understand the opportunities and risks of AI for children. With a Ph.D. in Cognitive Neuroscience and a Master’s Degree in Psychology, her research focuses on how AI intersects with child cognitive and emotional development.
• Munmun De Choudhury, Ph.D.—J. Z. Liang Professor of Interactive Computing at Georgia Tech. She harnesses computational approaches to better understand the role of online technologies in shaping and improving mental health.
• Tracy Dennis-Tiwary, Ph.D.—Professor of Psychology at Hunter College and co-founder and CSO at Arcade Therapeutics. She creates digital games for mental health and explores interactions between technology and emotional wellbeing. 
• Sara Johansen, M.D.—Clinical Assistant Professor at Stanford University and founder of Stanford’s Digital Mental Health Clinic. Her work explores how digital platforms can support mental health and well-being.
• David Mohr, Ph.D.—Professor at Northwestern University and Director of the Center for Behavioral Intervention Technologies. He studies how technology can help prevent and treat common mental health conditions such as depression and anxiety.
• Andrew K. Przybylski, Ph.D.—Professor of Human Behaviour and Technology at the University of Oxford. He studies how social media and video games shape motivation and well-being.
• Robert K. Ross, M.D.—A national leader and expert in health philanthropy, public health, and community-based health initiatives. He began his career as a pediatrician and is formerly the president and CEO of The California Endowment. 

How we will work with the council members 

We kicked off this council last week with an in-person session to deep dive into OpenAI’s current work in these areas and to have council members meet the teams they will be working with and advising. 

Our work with the council will include regular check-ins on our approach, and recurring meetings to explore topics like how AI should behave in complex or sensitive situations and what kinds of guardrails can best support people using ChatGPT. As an example, for the rollout of parental controls, we consulted individual members to help us prioritize which controls to build first and how best to notify parents if their teen appears to be in distress. Their feedback shaped the tone of the messages we use, so they feel caring and respectful to both teens and the family members. 

The council will also help us think about how ChatGPT can have a positive impact on people’s lives and contribute to their well-being. Some of our initial discussions have focused around what constitutes well-being and the ways ChatGPT might empower people as they navigate all aspects of their life. We’ll keep listening, learning, and sharing what comes out of this work.

Expanding our safety work 

Alongside the Expert Council on Well-Being and AI advising on our broader approach to well-being, we’re also working with a multidisciplinary subset of mental health clinicians and researchers within the Global Physician Network to shape our model behavior and policies, and to test how ChatGPT responds in real-world situations. This work spans psychiatry, psychology, pediatrics, and crisis intervention, helping ensure our systems are grounded in clinical understanding and best practices.

We’ll have more to share soon about the improvements underway⁠ to the main ChatGPT model and what we’re learning about how it can best support people.

Generated by RSStT. The copyright belongs to the original author.

Source