The AI Advantage: Developing Trusted, Ethical & Accessible AI Augmented Human Decision Making & Automation for SMBs

人工智能正在许多领域掀起革命。然而，随着人工智能的使用越来越多，人们越来越担心人工智能驱动的决策过程的可靠性、风险和伦理，以及它可能导致的潜在危害。因此，制定标准、解决方案和流程以提高对人工智能系统的信任和可靠性、其合规性和风险管理至关重要。Humble AI正在开发一种人工智能系统，可以跟踪人类的决策，显示历史决策，对高度自信的决策进行分类，并实现端到端流程的自动化。该系统将允许更快和更可靠的决策，并将为分类决策提供有效的途径，而不会影响保证级别或风险缓解。此外，该系统允许影响评估和评价，并不断提高其可靠性和鲁棒性。该项目的目标是：a)开发一种人工智能系统，可以跟踪人类的决策并显示类似项目的历史决策；b)随着时间的推移，对高度自信的决策进行分类，从而提高人工智能系统的准确性和可靠性；c)自动化端到端流程，从而减少对人工干预的需求，提高效率——同时保持对人类系统的可听性和问责制。该项目将通过在决策过程中提供透明度、问责制和可靠性，在人工智能保障领域提供明确的、改变游戏规则的干预措施。该系统还将允许审计和影响评估或评价，提高对人工智能及其持续发展的信任和合规性。

AI is revolutionising many areas. However, with the increasing use of AI, there are growing concerns surrounding the reliability, risks and ethics of AI powered decision-making processes, and the potential harm it can lead to. Therefore, the development of standards, solutions and processes to advance trust in and reliability of AI systems, their compliance and risk management, is essential. Humble AI is developing an AI system that can track human decisions, display historic decisions, triage highly confident decisions, and automate the end-to-end process. The system will allow for faster and more reliable decision making, and will provide an efficient route for triaging decisions, without impacting on assurance levels or risk mitigation. Furthermore, this system allows for impact assessment and evaluation, and continuously improves its reliability and robustness. The project will aim to a) develop an AI system which can track human decisions and display historic decisions for similar items b) triage highly confident decisions over time, thus improving the accuracy and reliability of the AI system c) automate the end-to-end process, thus reducing the need for human intervention and improving efficiency - whilst maintaining audibility and accountability for the human system. This project will deliver a clear, game-changing intervention in the field of AI assurance by providing transparency, accountability, and reliability in the decision-making process. The system will also allow for auditing and impact assessment or evaluation, improving the trust in and compliance of AI and its ongoing development.