Developing, Evaluating and Commercialising Scalable Automated Multiple Mini Interviews

大型跨国公司平均花费4000 GB招聘一名新员工；每年要填补4000多个职位空缺，成本可能超过1600万GB。面试官的偏见在各个行业都很普遍；根据英国政府的数据，在企业最高领导层，96.7%的人是白人，71%是男性。这个由Innovate UK资助的项目为期18个月，将使一种最先进的自动候选人选择系统SAMMI(可伸缩自动多迷你面试)得到改进、测试并投放市场，旨在让合适的人进入合适的工作岗位，同时提高成本效益和多样性招聘。SAMMI目前是一个原型，旨在提供可扩展的、英国健康教育认可的面对面多次迷你面试(MMI)技术的自动化版本，广泛用于医疗保健学生的选择。MMI要求考生坐在一系列旨在评估软技能的、无偏见的“真实生活场景”中。Sammi建立在十多年的MMI研究和文本分析的基础上，拥有一个定制的在线系统。对话分析技术评估申请者的面试回应(文本/演讲)，人工智能能够根据预定义的软技能对候选人的答案/互动进行评分。对应聘者软技能的分析提取适用于：预筛选、选择和与简历/个人陈述/求职信的比较。Sammi原型于2018年在萨里大学开发。它已经在曼彻斯特医学院和萨里大学进行了测试，证明了良好的功能和测试重新测试的可靠性。Innovate UK资助的市场验证之旅显示，100多家公司/组织(私人、公共和慈善机构)对Sammi产生了广泛的国际兴趣，这些公司/组织正在进行大规模的候选人选择。该项目将利用Sammi更快地进入市场，从而利用先前的投资和建立的客户联系。将开发、改进和测试SAMI的核心组成部分(以确保无偏见)，并计划对SAMI的有效性(可靠性、有效性、功能性、可接受性和成本效益与面对面比较)进行案例研究评估。与潜在试点地点的讨论正在进行中，例如：英国和美国医学院；国际非政府组织和跨国公司。SAMMI将利用时间和资源来增强SAMI，提供补充的‘附加组件’，包括一个新的申请人跟踪系统，将简历/个人陈述内容与面试表现进行比较。SAMMI提供了一个独特的系统，旨在降低成本效益，并避免无意的面试偏见。

Large multinational companies spend, on average, £4k hiring each new employee; with over 4000 vacancies to fill a year, the costs can exceed £16m. Interviewer bias is widespread across sectors; according to UK government figures, at the most senior level of corporate leadership, 96.7% of individuals are white and 71% male.This 18-month, project, funded by Innovate UK, will enable the refinement, testing and launch to market of a state-of-the-art automated candidate selection system, SAMMI (scalable automated multiple mini interview) designed to get the right people into the right job while increasing cost-efficiency and diversity hiring.SAMMI is currently a prototype designed to offer a scalable, automated version of the face-to-face 'Multiple Mini Interview' (MMI) technique endorsed by Health Education England and widely used in healthcare student selection. MMIs involve candidates sitting a series of 'stations' which present bias-free, 'real life scenarios' designed to assess soft skills. SAMMI builds on over ten years of MMI research and text analytics with a custom built online system. Conversational analysis techniques evaluate applicants' interview responses (text/speech) and artificial intelligence enables scoring of candidate answers/interactions against pre-defined soft-skills. The analytics extracted of candidates' soft skills are applicable to: pre-screening, selection and comparison against CVs/personal statements/cover letters.The SAMMI prototype was developed at the University of Surrey in 2018. It has been tested at Manchester Medical School and The University of Surrey demonstrating good functionality and test re-test reliability. The Innovate UK funded market validation journey revealed extensive international interest in SAMMI across 100+ companies/organisations (private, public and charity) undertaking large scale candidate selection.This project will leverage SAMMI to enable quicker market entry thereby capitalising on prior investment and established customer contact. Core SAMMI components will be developed, refined, tested (to ensure bias-free) and case study evaluation of the effectiveness (reliability, validity, functionality, acceptability and cost-benefit compared to a face-to-face comparator) of SAMMI is planned. Discussion are ongoing with potential pilot sites for example: UK and US Medical Schools; international non-governmental organisations and multinational companies. Time and resources will be used to enhance SAMMI with complementary 'add-ons' including a novel applicant tracking system comparing CV/personal statement content with interview performance.SAMMI offers a unique system designed to be cost-efficient as well as avoid unintentional interview bias.