Conference: Inaugural Workshop on Provably Safe and Beneficial AI (PSBAI)

本次研讨会提案旨在将广泛的研究人员聚集在一个非常相关和雄心勃勃的主题上。它将及时开始讨论如何使人工智能（AI）安全用于公共利益，以及该领域的当前进展如何成为该应用的支持骨干。该提案涉及的主题具有很大的研究潜力，但迄今为止在工程和计算研究社区中很少有人访问。总的来说，这是一个强有力的研讨会提案，将导致关于当今社会两个重要主题的新颖研究对话-人工智能和数字生态系统。该研讨会将在加州大学伯克利分校校园举行，为期两天，包括面对面和视频会议，有50名与会者，包括来自行业，非营利组织，学术界，政府和超国家机构的利益相关者，以欧盟委员会，世界经济论坛，经合组织，联合国教科文组织和GPAI多年的合作和咨询为基础；与Schmidt Futures、Kavli、Open Philanthropy等主要基金会合作；还有几个国家的政府。PI做得很好，汇集了一个令人印象深刻的议程，包括学术界，工业界和政府组织。提议的研讨会有五个建议的主题领域，包括：1)安全和有益的人工智能系统的基本主题；2)对可证明的安全和有益的算法媒体的考虑；3)可证明的安全和有益的人机交互的考虑；4)相关的技术考虑，以实现在安全的数字生态系统中管理强大人工智能系统部署的监管框架；5)关于公平准入的考虑。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

This workshop proposal aims to bring together a wide range of researchers on a very relevant and ambitious topic. It will start a timely discussion on how Artificial Intelligence (AI) can be made safe for public good, and how current advances in the field can become a supporting backbone for that usage. The proposal addresses a topic that has a lot of potential for research but has been very sparsely visited so far in both engineering and computing research communities. Overall, this is a strong workshop proposal which will lead to novel research conversations on two major topics of importance for today’s society – AI and digital ecosystem. The workshop is to be conducted over 2 days at the University of California Berkeley campus, both in-person and via videoconference, with 50 attendees, including stakeholders across industry, non-profits, academia, government and supra-national institutions, building on years of collaboration and consultation with the EU Commission, World Economic Forum, OECD, UNESCO, and GPAI; and with major foundations including Schmidt Futures, Kavli, Open Philanthropy; and several national governments. The PI has done a good job in putting together an impressive agenda, featuring academia, industry, and government organizations. The proposed workshop has five suggested topic areas including: 1) fundamental topics for safe and beneficial AI systems; 2) Considerations for provably safe and beneficial algorithmic media; 3) Considerations for provably safe and beneficial human-robot interaction; 4) Relevant technical considerations to enable a regulatory framework governing the deployment of powerful AI systems within a secure digital ecosystem; and 5) Considerations regarding fair access.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.