Profiting from science: the development and valuation of Artificial Intelligence new ventures and engagement with the science base

背景：英国政府寻求将英国定位为人工智能(AI)领域的领导者。在这种背景下，人工智能通常被视为一种新的变革性通用技术(GPT)。然而，我们缺乏证据和理解，不知道人工智能技术是如何被新企业用来创造和获取价值的。与人工智能一样，新的GPT可以重塑行业动态，导致新产品、服务和商业模式的开发。在克服GPT新兴技术的监管、市场和技术不确定性方面，以科学为基础的新企业往往发挥着关键作用。然而，人工智能的特点是其学术根源和应用驱动的软件开发之间的双重性质，这使得它的发展道路可能不同于最近的其他GPT，如生物技术和软件。此外，“以科学为基础”的类别并没有涵盖人工智能公司正在采用的更广泛的科学参与方法。事实上，一些人工智能新企业似乎是科学界的领军人物，而另一些企业则在相反的极端，与科学很少或根本没有互动。目标：目前，我们对人工智能新企业如何与研究基地互动并从中获利(或不获利)的了解不大。这项提议的目的是在理解这个问题上开辟新的天地。它借鉴了三个紧密结合的项目，以产生对新企业如何从科学中获利的新见解。首先，它将导致一个新的人工智能风险和研究天文台(AIVRO)的发展，该天文台将揭示英国的人工智能新企业及其与研究基地的接触。这项研究将通过收集来自公司报告、科学出版物、会议记录、研究合作和其他来源的数据，对新企业的科学参与进行分类。该数据集将是绘制英国人工智能创新生态系统的独特资源，为其他研究人员提供重要的研究工具。其次，该项目将评估科学投入对新企业的商业化战略、创新产出和业务绩效的影响，为与研究基地的互动可能塑造企业增长和发展提供直接证据。第三，通过一个实验，这项研究将因果地评估潜在投资者对人工智能新企业的科学参与所带来的价值。预期收益：这项研究将提供强有力的证据，为有关人工智能和英国研究基地的发展的管理和政策提供信息和支持。它还将扩大我们对新企业如何动员外部资源，特别是在科学系统内的资源，以帮助它们克服面临的挑战，并帮助获得外部参与者的支持的理解。新的数据库将使其他研究人员能够建立和扩展这一分析，以捕捉人工智能新企业的新的和不同的方面。我们还计划加强研究社区。我们将通过创建人工智能企业及其与研究基地的接触的新数据集、早期职业研究人员研讨会、关键会议的研讨会以及在领先期刊上发表文章来做到这一点。此外，该项目将通过我们的工业和政策咨询委员会以及政策研讨会和报告吸引广泛的从业者。我们的Avrio数据库还将提供一个用户友好的机会，用于分析和可视化英国人工智能生态系统中研究和公司之间的联系。

Context: The UK government seeks to position the UK as a leader in Artificial Intelligence (AI). In this context, AI is often seen a new transformative, General-Purpose Technology (GPT). However, we lack evidence and understanding of how AI technologies are being used by new ventures to create and capture value. Novel GPTs, like AI, can reshape industries' dynamics, leading to the development of new products, services, and business models. Science-based new ventures often play a critical role in overcoming regulatory, market, and technical uncertainty in emerging GPT technologies. Yet, AI is characterized by a dual nature between its academic roots and application-driven software development, making its development pathway potentially different from other recent GPTs, such as biotechnology and software. Moreover, the "science-based" category does not capture the broader spectrum of scientific engagement approaches that AI firms are adopting. Indeed, some AI new ventures appear to be leading actors in scientific communities, while others, at an opposite extreme, have little or no interaction with science. Objectives: At present, our understanding of how AI new ventures interact with and profit (or not) from the research base is modest. The objective of this proposal is to break new ground in understanding this question. It draws on three tightly integrated projects to generate new insights into how new ventures profit from science. First, it will lead to the development of a new AI Venture and Research Observatory (AIVRO) that will reveal the universe of AI new ventures in the UK and their engagements with the research base. The research will develop a taxonomy of new ventures' scientific engagement by drawing together data from company reports, scientific publications, conference proceedings, research collaborations and other sources. This dataset will be a unique resource to map the UK AI innovation eco-system, offering a significant research tool for other researchers. Second, the project will assess the impact of scientific engagement on new ventures' commercialisation strategies, innovative outputs, and business performance, providing direct evidence about the interactions with the research base may shape venture growth and development. Third, using an experiment, the research will causally assess the value ascribed to scientific engagement for AI new ventures by potential investors. Expected benefits: The research will provide robust evidence to inform and underpin management and policy concerning the development of AI and the research base in the UK. It will also extend our understanding of how new ventures mobilise external resources, especially within the science system, to help them overcome the challenges they face and to help garner support for external actors. The new database will enable other researchers to build and extend this analysis to capture new and different aspects of AI new ventures. We also plan to strengthen the research community. We will do so through the creation of a new dataset of AI ventures and their engagement to the research base, an Early Careers Researchers workshop, symposia at key conferences, and publications in leading journals. In addition, the project will engage a broad range of practitioners, through our Industry and Policy Advisory Board, and policy workshops and reports. Our AVRIO database will also offer a user-friendly opportunity for analysis and visualisation of the links between research and companies in the UK AI eco-system.