Ai Risk Management for SME ecosystem

中小企业生态系统的人工智能风险管理

• 批准号: 10045611
• 金额: $ 40.04万
• 依托单位: HOLISTIC AI LIMITED
• 依托单位国家: 英国
• 项目类别: Collaborative R&D
• 财政年份: 2023
• 资助国家: 英国
• 起止时间: 2023 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=10045611
• 关键词: Ai; Risk; Management; SME; ecosystem

Until recently, the use of AI in business has been generally unregulated. However, this is rapidly changing. AI's increasing prevalence and wider consumer and legislator recognition of its role in society is driving forward a swathe of legislation across the globe. The UK has recently announced its National AI Strategy which includes the Centre for Data Ethics and Innovation AI 'assurance ecosystem roadmap' and the FCA, ICO, OFCOM and CMA are all producing regulatory guidance. Internationally, EU AI Acts and multiple US interventions are confirmed as emerging over the next year. Most notably, New York City now requires 'bias audits' of all AI-HR technologies.Holistic AI ensures that enterprises can adopt AI with confidence and assurance guaranteeing AI that is safe, ethical, and legal. This is achieved by intensively assessing data, models, and processes, at any level of maturity or scale.We will create an SME-focussed evaluation tool to enable them to access our Enterprise-grade AI assurance system. 432,000 UK companies have adopted at least one AI technology (DCMS: AI activity in UK businesses-2022) and these SMEs must adapt to the fundamental restructuring of our data economy, now and for the longer-term. We will enable them to adopt AI with confidence, minimising risk, and prepare for demonstrating compliance with, for example, the wide-ranging implications of the EU AI Act.This is a significant challenge as it requires creating a far more streamlined version of the current platform with far less computationally expensive assessments, more in line with potentially simpler use-cases and ecosystems. Our current deployments with enterprise organisations spans systems from Windows Hello through recruitment to intensive research platforms, with analysis conducted over multiple stages spanning Risk Mapping, Risk Mitigation, Assurance Reporting and Bespoke Advisory.

直到最近，人工智能在商业中的使用一直不受监管。然而，这种情况正在迅速改变。人工智能的日益普及以及消费者和立法者对其在社会中作用的更广泛认可，正在推动地球仪的立法。英国最近宣布了其国家人工智能战略，其中包括数据道德和创新中心人工智能“保证生态系统路线图”，FCA，ICO，OFCOM和CMA都在制定监管指南。在国际上，欧盟人工智能法案和美国的多项干预措施被证实将在明年出现。最值得注意的是，纽约市现在要求对所有人工智能-人力资源技术进行“偏见审计”。整体人工智能确保企业可以有信心和保证地采用人工智能，保证人工智能是安全的，道德的，法律的。这是通过对任何成熟度或规模的数据、模型和流程进行深入评估来实现的。我们将创建一个以中小企业为重点的评估工具，使他们能够访问我们的高级人工智能保证系统。432，000家英国公司已经采用了至少一种人工智能技术（DCMS：英国企业的人工智能活动-2022年），这些中小企业必须适应我们数据经济的根本性重组，现在和长期。我们将使他们能够充满信心地采用人工智能，最大限度地降低风险，并准备证明符合欧盟人工智能法案的广泛影响。这是一个重大挑战，因为它需要创建一个更精简的版本，以更少的计算成本进行评估，更符合潜在的更简单的用例和生态系统。我们目前与企业组织的部署涵盖了从Windows Hello到招聘再到密集研究平台的系统，并在多个阶段进行分析，包括风险映射、风险缓解、保证报告和定制咨询。