Artificial intelligence (AI) is used in many areas such as credit scoring. AI may also open up new possibilities in social work, for instance in the context of service planning for people with disabilities. However, it is frequently the case that algorithmic-based decisions are not necessarily transparent and that the criteria upon which such decisions are based are not readily understood by humans.

‘If people were to rely heavily on algorithmic recommendations during their decision-making process, it could undermine fundamental democratic principles.,’ says Dr. Philip Cimiano, a professor at Bielefeld University’s Faculty of Technology and the Center for Cognitive Interaction Technology (CITEC). ‘At the same time, algorithmic assistance in decision-making also offers the possibility of improving the quality of decisions. In this case, decisions are based on a much larger set of data and can, therefore, be more objective than human decisions.’ With both of these considerations in mind, researchers from CITEC and the University of Applied Sciences Bielefeld are laying the foundation to optimize algorithm-based assistance for decision-making.

To be able to support decision-making, researchers have to feed data into the algorithms. The v. Bodelschwingh Foundation Bethel, the largest welfare institution in Europe, is providing Professor Philipp Cimiano and his colleague Angelika Maier with data for their research. Bethel social institutions include residential services such as individual apartments or group-homes for disabled individuals.  

The data used in the project are drawn from Bethel case records on individuals who are part of a program that works to guarantee and promote their social participation – so called  “service planning .” Using algorithms, patterns can be identified in this service planning data. These patterns offer insight into the nature of service planning and the specific issues that arise during the process. Both become the basis for decision-making supported by intelligent systems. ‘We consider which issues a client is facing over a period of time, or how their autonomy is assessed,’ says Angelika Maier. ‘With this information, an individual’s needs and resources can be met in a more targeted way.’

‘The new system can help professionals at social institutions to provide their clients with even more targeted support in the future,’ says Professor Dr. Ingmar Steinhart, who is a member of the Board of Management of the v. Bodelschwingh Foundation Bethel.

Introducing new technologies like decision support systems brings up many issues. Professor Dr. Udo Seelmeyer and doctoral researcher Diana Schneider from the Faculty of Social Services at the Bielefeld University of Applied Sciences are investigating what needs to be considered when introducing a new system and where it can be usefully applied within an institution and in the decision-making process. In her doctoral research, Schneider deals with methods of scenario analysis and technology assessment. ‘What unintended side effects and unwanted social consequences might such a technical system have? How should it be used to enhance the quality of care and professionalism of the workers?’ asks Diana Schneider. Udo Seelmeyer also sees an advantage for the professional work: ‘A decision support system offers new points of access to the vast mass of textual documentation that otherwise remains unused in the data graveyard,’ says Seelmeyer.

Algorithms may be able to make recommendations for decisions, but they will not replace the expertise and implicit knowledge gleaned from experience that social workers have. ‘Our prototype is meant to ensure – and not undermine – basic democratic principles of inclusion, participation, and equality,’ explains Cimiano. ‘It is our aim to create a clear vision of how a system providing algorithm-based recommendations for decision-making can be used for the benefit of everyone.’

The project is called “Maewin” and it will run through mid 2021. “Maewin” is the German acronym for “Maschinelle Entscheidungsunterstützung in wohlfahrtsstaatlichen Institutionen“ [Decision-Support Systems in welfare institutions]. This research is supported by the Digital Society research program funded by the Ministry of Culture and Science of the German State of North Rhine-Westphalia.. The Digital Society research program addresses questions of how digitalization is transforming society and how technology impacts the formation of opinions.

More information is available online:

• Info sheet on the Maewin project (in German)
• Maewin profile on the homepage of the Digital Society research program (in German)

Contact:
Prof. Dr. Philipp Cimiano, Bielefeld University
Faculty of Technology
Phone: +49 521-106 12249
Email: cimiano@techfak.uni-bielefeld.de