Team of researchers at FAU develop adaptive assistant for deleting superfluous files
Ever larger numbers of files are stored in the administrative departments of companies and it’s sometimes difficult for staff to keep on top of the flood of spreadsheets, presentations and documents. In addition, each file uses energy and uses up storage space on a hard drive, creating costs for the company and damaging the environment. But which files can be deleted? Where is the best place to start?
Psychologists at FAU are currently working with computer scientists from the University of Bamberg to find an automated solution. They are developing an adaptive system that cooperates with users using machine learning in an interdisciplinary project called ‘Dare2Del’ that helps to find and delete irrelevant data. Psychologists Prof. Dr. Cornelia Niessen and Kyra Göbel at the Chair of Work and Organizational Psychology at FAU have discovered that people only rarely delete files at work, and mostly only when these files interfere with their work. The aim of the new system is to increase people’s willingness to delete files, help them to focus and improve their concentration.
During the first phase of the project from 2016 to 2019, Prof. Dr. Ute Schmid and Michael Siebers from the University of Bamberg programmed an adaptive system that can process extremely complex algorithms but is still easy to use. The new system does not delete files at random and certainly not automatically either. The artificial intelligence used in the system takes company regulations and legal requirements into account and adapts to the needs of the users. For example, they can define rules such as that the latest and penultimate version of a file must not be deleted. A particular challenge for the researchers during the project is finding out when users want to use which contents. Someone in the middle of an important project doesn’t want to be asked if they still need a file from a completely different context.
‘From a psychological perspective, the plausibility of decisions made by the system is a decisive factor in ensuring that users can trust the system’
During the next phase of the project in the next three years, the researchers intend to make sure that the system’s suggestions are plausible and clear for users. For example, the programme should highlight specific file names and display explanations about why it is suggesting a file should be deleted. ‘From a psychological perspective, the plausibility of decisions made by the system is a decisive factor in ensuring that users can trust the system and not worry about the decision later, but simply forget about the file’, says Prof. Niessen. In addition, the project is also investigating who would set to benefit the most from a system like ‘Dare2Del’ and thus improve how they manage the flood of information.
The ‘Dare2Del’ project is set to run for six years and has received around 890,000 euros of funding from the German Research Foundation (DFG). It is part of the DFG Priority Programme called ‘Intentional Forgetting in Organisations’.
Prof. Dr. Cornelia Niessen
Phone: + 49 9131 85 64001