About the project

'Do Algorithms Care?' is an interdisciplinary collaboration between artist Amanda Bennetts (AU) and data scientist Johanna Einsiedler (AT). At the core of this project is an N=1 study, typically challenged in scientific circles for its limited generalizability, yet here serves as a critical methodological pivot. Using a DIY smartwatch and EMG (electromyography) and MMG (mechanomyogram) sensor kits, the duo engage in self-monitoring to collect their physiological data. This personalized data collection reflects a methodological shift in research practices, emphasizing the need for individual agency to manage and use biodata. The sensor measurements, smartwatch data and results from a wellbeing questionnaire are fed into a machine-learning model, developed to predict subjective fatigue.

Want to stay updated?


Project Beginnings

Amanda launched the project as a FOUNDING LAB Fall term project to explore AI's potential in managing her recently diagnosed rare muscle disease. By collecting and analyzing her own biodata using wearable technology, she aimed to gain specific insights that could help her better understand and manage her condition. This approach allowed her to combine artistic and scientific methods, questioning the effectiveness and ethics of AI in healthcare and wellness industries. Additionally, Amanda was interested in whether self-tracking offers real control over one's health or if it paradoxically limits it, exploring if it expands bodily knowledge or confines it within data metrics.

Phase One: Fall Term

During the Fall Term, Amanda collaborated with a data science student from Johannes Kepler University (JKU) to develop an AI model predicting her need for rest days. Which involved using AI to analyze Amanda's unique physiological data, employing wearable technology equipped with EMG (electromyography) sensor kits and a daily subjective health questionnaire. The model initially reported 100% accuracy, which Johanna later discovered turned out to be due to a critical error, where the outcome variable had been mistakenly used as a predictive feature. This error, while not uncommon in data science, led the pair to discuss and question the reliability of algorithms and the potential for biases in data science. This inquiry led to the interdisciplinary underpinnings of Amanda and Johanna’s collaborative Spring Term project.

Phase Two: Spring Term

In the Spring Term, Johanna Einsiedler officially joined the project. Johanna’s addition marked a turning point in the project, where her expertise in social data science assisted in the research and development of a protocol for the data acquisition, data handling, and algorithm development to ensure transparency and fairness in AI applications. Johanna was drawn to the project not only because of her expertise but also due to her shared interests with Amanda in data visualization. She contributed significantly by analyzing the data and enhancing the machine learning model.

Amanda's background in practice based artistic research lent to further questioning of the key theme of the project—examining how embodied personal biodata and AI can be used in healthcare but also advocating for ethical practices that ensure algorithms are as unbiased and accurate as possible. This approach underlined the necessity of maintaining a critical perspective on the tools and methods used in data science, especially when applied to sensitive areas like personal health.

Together, they focused on:

By incorporating Johanna’s biological data, the project now compares and visualizes the collective dataset of both Amanda and Johanna. This expansion not only serves as an interdisciplinary mode of inquiry but also represents a methodological shift in research practices. It facilitates a fruitful dialogue on art and science collaborations, allowing them to communicate and gain insights into each other's specialist fields, thereby enriching the project's overall understanding and outcomes.

Broader Impact

Their collaborative effort highlights the need for rigorous scientific methods and ethical considerations in AI healthcare applications. Amanda and Johanna's work underscores the importance of critical analysis to ensure AI tools are both effective, accessible and equitable, ultimately aiming to empower individuals in managing their health data. By working with patients to create patient-driven healthcare, the project emphasizes learning from individuals' lived experiences rather than solely relying on generalized data sets. This approach aims to empower individuals in managing their health, demonstrating that science can gain valuable insights from personalized data and real-world experiences, thereby fostering more nuanced and effective healthcare solutions.

The project is part of the Founding Lab of the Interdisciplinary Transformation University Linz.

About the people

Amanda Bennetts is an Australian new media and installation artist. Living with a progressive neurological disease, Bennetts draws on her experience to critically dissect issues relating to care, sickness and disability. Examining what it means to be an ill-body in the world, Bennetts navigates a realm of living that is politically charged and socially determined. She creates large installations that engage materiality, incorporating video, sound and mass-produced objects, often with clinical and disability aesthetics.

Johanna Einsiedler is a PhD student in Social Data Science at the University of Copenhagen. Her primary area of research are applications of network science methods and machine learning algorithms in the Social Sciences. Through her collaboration with artist Amanda Bennetts she wants to explore if and how algorithms can be used to improve an individual’s wellbeing and advocate for more DIY technology development.