The KI.ASSIST project - Assistance Services and Artificial Intelligence for People with Severe Disabilities in Vocational Rehabilitation - funded by the German Federal Ministry of Labour and Social Affairs (BMAS), started in April 2019 and will run until March 2022.
The project consortium consists of
- the Federal Association of Vocational Rehabilitation Centres (BV BFW) as applicant for the Netzwerk berufliche Rehabilitation (NbR) and project coordinator,
- the Federal Association of Vocational Rehabilitation Centres for young adults with disabilities (BAG BBW),
- the Federal Association of Sheltered Workshops (BAG WfbM)
- as well as the German Research Center for Artificial Intelligence (DFKI).
You can find out more about the organizations involved in the project here: Project team (available in German)
For the first time, the project partners will systematically investigate in a scientifically substantiated and practice-oriented manner which groups of people at which learning and working places can sustainably benefit from the use of AI-based assistance systems. Thereby, people with their needs are in the center of the project.
The following figure shows the various key working areas and work packages in the project. At the same time, these also represent the various perspectives on the use and potential of artificial intelligence in vocational rehabilitation united in the project.
The Project Structure - Clusters and Work Packages
The KI.ASSIST project places people with disabilities at the center and at the beginning of all considerations regarding the use of AI-based assistance systems in vocational rehabilitation. Focusing on people means that we address a number of questions on the following topics in the person-centeredness cluster.
Ongoing digitalisation and the transformation of the educational landscape as well as the working environment that it is driving require new knowledge and new skills. This knowledge and the associated skills and abilities are the prerequisite for successfully participating in education, completing training programs and succeeding in the labor market. There is a broad consensus in the education discourse that "digital competencies" must be accorded a particularly high priority. "Digital skills" are by no means to be seen as an alternative to the "core skills" that have been at the center of attention up to now, such as social skills or self-competence. Rather, they are seen as supplements to be able to successfully take on the challenges of the 21st century.
In the project, digital competencies are promoted in two ways:
- The specialists and participants of the rehabilitation measures will be trained on site in the learning and experimental rooms. Modularized training modules are being developed for this purpose. In the first part of the training courses, the vocabulary revolving around the topics of " digitalisation" and "artificial intelligence" is taught and tested in a practical and internally differentiated manner.
- In a second step, application-specific training modules are developed that enable the LER participants to acquire the technology to be tested at the respective location, to put it into use and to learn about its possible applications.
The project will be supported in the design and delivery of the training by an external agency specialised in the development of digitally-enabled and inclusive education. The training modules as well as the trainings themselves will be comprehensively evaluated to ensure learning success. Recommendations for action will be developed for vocational rehabilitation institutions for people with disabilities.
In the KI.ASSIST project, the topic of data sovereignty is placed in the context of the use of AI-based assistance systems by people with disabilities. The goal is to create guidelines and recommendations for action on questions of informational self-determination and data sovereignty of people with different impairments and disabilities when using AI-based assistance services. These are intended to educate and raise awareness for a sovereign use of AI technologies. For this purpose, studies and recommendations on the topic area will be evaluated on the one hand. On the other hand, interviews will be conducted with data protection experts as well as people with different types of disabilities in order to ascertain the level of knowledge as well as needs and, based on this, to better assess possibilities for training and awareness-raising measures.
Since the issue of data sovereignty is not only the responsibility of the end user, but should also be considered when selecting technologies, the work package will examine potentially suitable AI-based assistance technologies for their supporting role. Technologies that are data-saving and give users as much control as possible over their own data, for instance, are proving to be strongly supportive technologies for the sovereign use of AI-based assistance systems.
Acceptance and motivation
The work package deals with the acceptance of AI-based assistance technologies by users. The latter are on the one hand the people with disabilities in the institutions of vocational rehabilitation, who test the application in the learning and experimentation room. But also, the professionals of the facilities come into play when it comes to the questions under which conditions AI-based technologies are accepted when they are introduced, and how AI-based technologies can contribute to a motivating work environment.
We approach these questions in several ways:
- Through online surveys and interviews with professionals and people with disabilities during on-site visits to facilities.
- In addition to the interviews, the so-called nueva evaluation will be conducted in all innovation spaces. Nueva (Nutzerinnen und Nutzer evaluieren = users evaluate) is a separate evaluation model to determine the quality of a service. It is aimed at people with disabilities and is characterised by the fact that the evaluators themselves have learning difficulties or disabilities. They thus work as experts in their own right. The nueva evaluation is carried out in the form of a standardised online survey and interviews.
- Existing scientific findings on motivation and acceptance in the context of work, AI, and disability are continuously incorporated into the project activities.
Self-determination, participation and diversity
The aim of the work package is to promote equal opportunities in all learning and experimental rooms under the guiding principle of diversity. In doing so, we attach particular importance to a participatory design of the learning and experimental rooms in order to be able to adequately take into account the sometimes very different needs of the heterogeneous target group of people with disabilities, but also of the different institutions.
The work package accompanies the establishment of the participation principle in the learning and experimental rooms as well as the implementation of equal opportunities in the application scenarios. In addition, relevant publications are evaluated in order to review legal aspects and the current state of research on the professional participation of people with disabilities.
The findings, together with the experiences from the learning and experimental rooms and the work results from a project-accompanying expert group on the topic of "Inclusive working world of the future", will flow into action guidelines for participation, self-determination and equal opportunities in the use of AI-based assistance technologies for people with disabilities in vocational rehabilitation and on the labor market. In this way, the work package also addresses the overall goal of the project, which is to support the path (back) to the general labor market with the help of digital assistance services and artificial intelligence applications, thereby promoting the self-determined participation of people with disabilities in life in society in the long term.
Within the Exploration cluster, we are taking AI-based assistance technologies to selected vocational rehabilitation facilities, including three vocational rehabilitation centres for young adults with disabilities, three vocational rehabilitation centres, three sheltered workshops, and one company. At a total of 10 locations, people with disabilities and skilled workers are testing some of the technologies identified in the monitoring subsection in so-called innovation spaces (in German: Lern- und Experimentierräume, LER for short). For this purpose, the needs and wishes of the people are first surveyed and implementation scenarios for the LER with suitable technologies are developed. Subsequently, the technologies are implemented on site and tried out by people with disabilities while learning and working.
The goal of the exploration is to explore, through on-site testing, how AI-based technologies can support people with disabilities in educational and work processes and overall participation in working life. However, we are also asking what competencies, what data protection rules and what prerequisites are required in the institutions, but also overall in vocational rehabilitation. It is also important for us to find out how people with disabilities and professionals accept these technologies and what opportunities and risks they see.
The exploration will be accompanied by an external evaluation. In the course of the project, our practical experience and our research findings will be incorporated into recommendations for action, which will be made available via the dialog platform to both vocational rehabilitation institutions and an interested specialist audience.
What AI-supported technologies are available and how suitable are they for people with disabilities when working and/or learning? This is the core question that the KI.ASSIST subsection "Monitoring" wants to answer in the course of the project.
The first step is to systematically search for and record technologies with AI components that support people with disabilities in their work and training. The basis for this research is a search, description and evaluation system developed in the project network.
The researched technologies are evaluated in the context of a qualitative content analysis. The aim is to provide an overview of the state of the art in the development and use of AI-based assistance technologies.
The second step is validation, i.e. selected technologies are evaluated with regard to their suitability for people with disabilities in the workplace or training place. For this purpose, a panel of about 100 inclusion and AI experts will be established. These experts will take part in various surveys, e.g., guided telephone interviews or an online survey. Here, the experts use their own experience to evaluate the suitability of technologies presented in the form of profiles, discuss factors influencing the successful introduction of AI-based assistance technologies, and discuss problems and trends to be expected.
The results of this evaluation in the second step are, on the one hand, a long and short list of potentially suitable AI-based technologies and, on the other hand, identified problem areas, trends and influencing factors for the introduction of such technologies in vocational rehabilitation facilities.
The third step is dedicated to answering another central research question in the monitoring cluster: which AI-based assistance technologies are most promising under which conditions or are most likely to become relevant in the future?
In the foresighting, selected panel participants develop scenarios of AI-based assistance technologies in vocational rehabilitation based on the results from the validation (interviews, online survey) in a workshop. These scenarios are subsequently evaluated and discussed in a real-time Delphi survey by the entire panel. The results are validated, probable scenarios of which AI-based technologies will be relevant and realistically applicable in vocational rehabilitation in the future.
In the Transformation subsection of the KI.ASSIST project addresses the question of how change processes are to be designed so that people with disabilities can benefit from AI-based assistance systems in vocational rehabilitation and at work. The goal is to develop a model for transformation processes and, based on this, recommendations for action for the development, introduction and long-term use of AI-based assistance systems for people with disabilities, even beyond the KI.ASSIST project.
The focus of the Transformation cluster is on change processes in the introduction and use of AI-based assistance systems in vocational rehabilitation facilities and companies, but also on the structures, processes and actors in vocational rehabilitation as well as on the political, ethical and legal framework conditions of the transformation.
To this end, on the one hand, existing studies and recommendations on transformation processes will be systematically researched and evaluated. On the other hand, interviews will be conducted with experts from research, associations, involved actors in vocational rehabilitation (e.g. rehabilitation providers, integration offices, vocational rehabilitation institutions) as well as companies. In addition, findings from other subsections of the project (e.g., monitoring, exploration, competencies) will be used to answer the research questions in the area of transformation. In addition, a group of experts is addressing ethical questions.
The National Dialogue Platform is an important component of the project. Its information, dialog and participation formats are intended to reach very different groups, from people with disabilities to experts from science and research to the broader public. The focus is on the exchange around the potentials of artificial intelligence for the design of an inclusive working world.
Externally, the KI.ASSIST website is a central element of the dialog platform. It provides access to the topics and activities in the project for all interested parties. The recommendations for action developed in the various subsections of the project are also published here, making it possible to transfer experience and knowledge to other vocational rehabilitation institutions. For the project teams of the learning and experimental rooms as well as the experts actively involved in the project, the internal area of the website also offers the possibility to organize themselves in working groups and to exchange experiences. In addition, the different phases of the conception and implementation of the learning and experimental rooms are accompanied by various - in part digital - workshops and participation offerings.
All results of the project will also be presented at symposia and a national summit as a final event. Here, they can be discussed with representatives of employers' associations, trade unions, science academies, professional policy, digital and IT associations, companies, and the Vocational Rehabilitation Network.