One aspect of a responsible application of Artificial Intelligence (AI) is ensuring that the operation and outputs of an AI system are understandable for non-technical users, who need to consider its recommendations in their decision making. The importance of explainable AI (XAI) is widely acknowledged; however, its practical implementation is not straightforward. In particular, it is still unclear what the requirements are of non-technical users from explanations, i.e. what makes an explanation meaningful. In this paper, we synthesize insights on meaningful explanations from a literature study and two use cases in the financial sector. We identified 30 components of meaningfulness in XAI literature. In addition, we report three themes associated with explanation needs that were central to the users in our use cases, but are not prominently described in literature: actionability, coherent narratives and context. Our results highlight the importance of narrowing the gap between theoretical and applied responsible AI.
MULTIFILE
The in-depth assessment of the situation of the European textile and clothing sector is composed by six independent reports with a close focus on key aspects useful to understand the dynamics and the development of the textile and clothing industry, drivers of change – most notably the impact of the financial crisis – and identification of policy responses and best practices. This has been done in six specific tasks leading to the six reports: Task 1 Survey on the situation of the EU textile and clothing sector Task 2 Report on research and development Task 3 Report on SME situation Task 4 Report on restructuring Task 5 Report on training and Education Task 6 Report on innovation practices. This final report draws on the key findings of each independent report, highlighting major conclusions in order to describe the situation of the textile and clothing industry and the way forward for the sector. In line with the terms of reference the findings in the six reports have been analysed in connection with the developments following the recommendations drafted by the High Level Group on textiles and clothing (further referred to as HLG), installed in 2004 as a response to the European Commission Communication of 29th of October 2003 on the textile and clothing industry. The HLG was composed of leading personalities representing stakeholders in the textiles and clothing industry and issued two reports entailing a vision and recommendations.
MULTIFILE
One aspect of a responsible application of Artificial Intelligence (AI) is ensuring that the operation and outputs of an AI system are understandable for non-technical users, who need to consider its recommendations in their decision making. The importance of explainable AI (XAI) is widely acknowledged; however, its practical implementation is not straightforward. In particular, it is still unclear what the requirements are of non-technical users from explanations, i.e. what makes an explanation meaningful. In this paper, we synthesize insights on meaningful explanations from a literature study and two use cases in the financial sector. We identified 30 components of meaningfulness in XAI literature. In addition, we report three themes associated with explanation needs that were central to the users in our use cases, but are not prominently described in literature: actionability, coherent narratives and context. Our results highlight the importance of narrowing the gap between theoretical and applied responsible AI.
MULTIFILE
This project aims to develop a measurement tool to assess the inclusivity of experiences for people with varying challenges and capabilities on the auditory spectrum. In doing so, we performed an in-depth exploration of scientific literature and findings from previous projects by Joint Projects. Based on this, we developed an initial conceptual model that focuses on sensory perception, emotion, cognition, and e[ort in relation to hearing and fatigue. Within, this model a visitor attraction is seen as an “experienscape” with four key elements: content, medium, context, and individual. In co-creative interviews with experts by experience with varying challenges on the auditory spectrum, they provided valuable insights that led to a significant expansion of this initial model. This was a relevant step, as in the scientific and professional literature, little is known about the leisure experiences of people with troubled hearing. For example, personal factors such as a person’s attitude toward their own hearing loss and the social dynamics within their group turned out to greatly influence the experience. The revised model was then applied in a case study at Apenheul, focusing on studying differences in experience of their gorilla presentation amongst people with varying challenges on the auditory spectrum.Societal issueThe Netherlands is one of the countries in Europe with the highest density of visitor attractions. Despite this abundance, many visitor attractions are not fully accessible to everyone, particularly to visitors with disabilities who sometimes are not eligible to ride due to safety concerns, yet when eligible generally still encounter numerous barriers. Accessibility of visitor attractions can be approached in various ways. However, because the focus often lies on operational and technical aspects (e.g., reducing stimuli at certain times of the day by turning o[ music, o[ering alternative wheelchair entrances), strategic and community-focused approaches are often overlooked. More importantly, there is also a lack of attention to the experience of visitors with disabilities. This becomes apparent from several studies from Joint Projects, where visitor attractions are being visited together with experts by experience with various disabilities. Nevertheless, experience is often being regarded as the 'core product' of the leisure sector. The right to meet, discover, develop, relax and thus enjoy this core product is hindered for many people with disabilities due to a lack of knowledge, inaccessibility (physical, digital, social, communicative as well as financial) and discrimination in society. Additionally, recreation entrepreneurs still face a significant gap in reaching the potential market of guests with disabilities and their networks. Thus, despite the numerous initiatives in the leisure sector aimed at improving accessibility on technical and operational fronts, often people with disabilities are still not being able to experience the same kind of enjoyment as those without. These observations form the pressing impetus for initiating the current research project, tapping into the numerous opportunities for learning, development and growth on making leisure offer more inclusive.Benefit to societyIn total, the current project approach comes with a number of enrichments in terms of both knowledge and methodology: a mixed-methods approach that allows for comparing data from different sources to obtain a more complete picture of the experience; a methodological co-design process that honours the 'nothing about us without us' principle; and benchmarking for a group (i.e., people with challenges on the auditory spectrum) that despite the size of its population has thus far mostly been overlooked.
Crowdfunding campaigns have empowered countless innovative projects and made funding accessible to a large pool of makers and citizens. Recently, traditional funding bodies such as foundations, provinces and municipalities have acknowledged the potential of crowdfunding to approximate institutional decision-making to citizens, by engaging with the “crowd’s” preferences and further stimulating public and private funding through matchfunding. Matchfunding – the financial contribution of traditional funding bodies to crowdfunding campaigns – is an emerging form of co-funding that has the potential to foster a more inclusive and democratic society. Yet, given its novelty, little is known about how matchfunding works, and how it can be transformed into an efficacious tool that supports project creators and policymakers to develop impactful projects. Looking at the creative industries, one of the most prolific fields in crowdfunding, this project aims to provide this knowledge by: (1) gaining insight into the democratizing potential and best-practices of matchfunding in the creative industries by comparing analysing the extensive databases of crowdfunding and matchfunding pioneer voordekunst and matchfunding partners Kunstloc Brabant and Gemeente Rotterdam, and by conducting interviews with matchfunding parners to gain insight into their challenges and experiences; (2) deepening and sharing findings in Impact-Driven Workshops, which serve to exchange knowledge with and between matchfunding partners, and gain further insights into their motives and best practices. Based on the outcomes of (1) and (2), we develop (3) an online Matchfunding Toolkit, geared towards matchfunding partners, as well as to creators, freelancers and SMEs (potentially) using matchfunding for their projects. Finally, (4) we will disseminate this knowledge to other funding bodies and organisations within and outside of the creative industries by connecting partners and stakeholders in a Dissemination Event. This results in a lasting knowledge hub and network geared to supporting creators, SMEs and freelancers in search of funding.
Toepassingen gebaseerd op artificiële intelligentie (AI) worden steeds vaker ingezet voor het maken van keuzes en besluiten. Deze toepassingen worden echter ook steeds complexer. Het is in sommige gevallen niet of moeilijk na te gaan hoe een algoritme tot een besluit is gekomen. Wat de AI doet is als het ware ondoorzichtig. Dit geldt ook in de financiële sector, terwijl juist in deze sector vertrouwen een grote rol speelt.
