ReRes

Student projects by Eva Krapp
Supervised project by Matthias Laschke & Robin Neuhaus | Siegen 2020


ReRes, a system for cognitive restructuring making use of the capabilities of non-human counterparts.


Technologies that approach us as a counterpart or, as the technology philosopher Don Ihde phrases it, as "quasi-other" have become increasingly prevalent. Amazon's Alexa and Google Home Assistant turn on our radiators, dim our lights, while Siri answers even the question for the meaning of life (it is, in fact, 42). In their design, all these assistants follow an anthropomorphic design, i.e., the technology should appear as human as possible. However, in this on-going Turing test, the unique qualities of technology and the many advantages of their 'Otherness' are getting lost.

The project ReRes – Cognitive Restructuring Otherware explores the idea of 'Otherness' and the possible advantages of machine-like qualities in a stress management context. It takes inspiration from worksheets often used as homework assignments in Cognitive Behavioural Therapy (CBT). CBT is a modern psychotherapy approach that operates on the central assumption that cognitions, behaviors, and emotions are interacting systems mutually influencing each other. Therefore, negative emotions and their consequences can be managed by actively reshaping thoughts and actions. CBT interventions are useful for alleviating symptoms of a variety of mood disorders and as a tool for managing everyday stress and increasing wellbeing.

One of the methods used in CBT is Cognitive Restructuring. In cognitive restructuring, the irrational thoughts which often accompany stressful situations are questioned and reshaped more helpfully. This can be achieved by identifying common cognitive distortions and disputing them from a rational standpoint.

ReRes explores a technological extension to in-person therapy's traditional modalities and worksheet exercises could look like – without being anthropomorphic. Instead of emulating and substituting a human interaction, ReRes emphasizes machine-innate qualities, such as being continually available, non-judgmental, endlessly patient, completely logical, and emotionally uninvolved. The idea is not to replace the conversation with a therapist, but to create a completely new situation.

The underlying metaphor which guided the design of ReRes is a debugging process, as used in software development. Like a debugging tool detects errors in faulty code, ReRes detects errors in faulty thought processes. As such, the analysis and restructuring of ReRes takes place solely on a semantic level, adhering to simple rules for detecting and replacing particular identifiers. Once the user enters a thought, ReRes checks it for identifiers indicating particular cognitive distortions and suggests an alternative phrasing.

While the functional and conceptual core of a cognitive restructuring tool was found relatively early in the project, creating a counterpart feeling without slipping into anthropomorphism proved quite the challenge. Therefore, ReRes evolved only through experimenting with several different designs and gradually homing in on what makes a counterpart feel like a counterpart. The design experiments ranged from a very technical command-line interface to a chatbot interface through a translator type interface.

Finally, ReRes was prototyped as a desktop application with distinct design elements evoking the feeling of interacting with a non-human counterpart. The interaction is structured in a conversation-like flow, visualized through a fluctuation of the user's and the agent's differently colored canvasses. The agent was given a life-like quality through animations such as bubbles signifying 'thinking' and 'marking' errors in the user's thought. These animations were adapted further to elicit an alive yet clearly technical and non-anthropomorphic perception in the latest design iteration.