Collaborative Research: FW-HTF-R: Toward an Ecosystem of Artificial-intelligence-powered Music Production (TEAMuP)

该项目为音乐生产建立了一个新的生态系统，使未来的音乐家能够更好地利用人工智能(AI)工具创作、表演和传播他们的音乐，同时还加快了音频AI研究。这包括创建一个开放获取的软件框架，使音乐家和研究人员能够在开发和使用越来越好的人工智能支持的音乐创作工具方面进行合作，以及一系列举措，使临界数量的音乐家能够以变革性的方式使用这些工具。预计音乐家将使用这些工具来生产成本更低、质量更高的音乐产品，以满足视频、网站、广告、录音和其他新媒体对数字音乐内容日益增长的需求。使音乐家在他们的音乐创作中更加自给自足，有可能增加有音乐天赋的个人的数量，他们将能够以他们的艺术谋生，特别是来自目前代表不足的人群。为了使成长中的音乐家能够充分利用人工智能工具，将开发一套创新的学习体验，以获得所需的心态和技能，并将在为期两个学期的课程中进行现场测试，该课程面向有音乐兴趣的学生，并为未被充分代表的大学预科青年举办“夏令营”，同时还将创建在线教学材料，以支持各种环境中的具体学习体验。该项目团队在音乐、音频工程、人工智能、学习科学/教育、商业/创业、伦理和包容性方面拥有互补的学科专业知识。这些技能将被用于为一个常用的免费和开源数字音频平台开发一个框架，该框架将允许：(A)音频AI研究人员可以轻松地将他们的新AI模型部署到该平台中；以及，b)使用这些AI工具与AI研究人员共享他们的音乐作品的音乐家，以便他们可以改进他们的模型。还将对不同的音乐家进行采访和调查，以更好地了解可能影响他们采用人工智能音乐制作工具的关键因素，这些工具可能如何改变他们的工作，以及大流行的影响和未被充分代表的人群在音乐制作中可能经历的其他障碍。总之，该项目将更好地了解可能影响音乐家在工作中采用和变革性使用人工智能的因素，了解可以推广到人类技术前沿的其他职业。最后，该团队将制定教学原则和实践，为有效的教育干预措施的设计提供信息，以更好地为未来的音乐家和其他领域的专家利用技术做好准备。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

This project builds the foundations of a new ecosystem for music production to empower future musicians to better leverage Artificial Intelligence (AI) tools in the creation, performance, and dissemination of their music, while also accelerating audio AI research. This involves the creation of both an open-access software framework enabling musicians and researchers to collaborate in the development and use of ever-better AI-powered tools for music creation, and a set of initiatives to enable a critical mass of musicians to use these tools in transformative ways. Musicians are expected to use these tools to produce lower-cost, higher-quality music products, which meet growing demand for digital music content for videos, websites, advertising, audio recordings, and other new media. Enabling musicians to be more self-sufficient in their music creation has the potential to increase the number of musically talented individuals that will be able to make a living with their art, especially from currently under-represented populations. To enable growing musicians to make full use of AI tools, a set of innovative learning experiences to acquire the needed mindsets and skills will be developed and field tested in a 2-semester course for students with music interests and a “Summer Camp” for pre-college under-represented youth, along with the creation of online instructional materials to support specific learning experiences in a variety of settings. The project team possesses complementary disciplinary expertise in music, audio-engineering, AI, learning sciences/ education, business/ entrepreneurship, ethics, and inclusion. These skills will be brought to bear on developing a framework for a commonly-used free and open-source digital audio platform that will allow: (a) audio AI researchers to easily deploy their new AI models into the platform; and, b) musicians who use these AI tools to share their music productions with AI researchers so they can refine their models. Interviews and surveys will also be conducted with diverse musicians to better understand key factors that may affect their adoption of AI music production tools and how those tools may transform their work, as well as the implications of the pandemic and other barriers that may be experienced by under-represented populations in music production. Together, the project will generate a better understanding of factors that may affect musicians’ adoption and transformative use of AI in their work, understanding which could be generalized to other occupations at the human-technology frontier. Finally, the team will develop pedagogical principles and practices that can inform the design of effective educational interventions to better prepare future musicians and other domain experts to leverage technology.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.