- But first, Tanja, you finished your work. What was it about? 

- In my thesis, I examined the increasingly complex knowledge infrastructures of voice, speech, and language technologies as these expand ubiquitously into everyday life. Well-known examples of the emerging “voice intelligence industry” include digital assistants (Siri, Alexa, Cortana), but the industry markets voice AI for applications much beyond human-computer interaction (HCI), says Tanja Knaus.
It is used for speaker identification, market analysis, and call center monitoring, and, more recently, voice AI is being marketed to the healthcare sector. In health care, the voice serves as a rich medium for diagnosing respiratory, mental, and neurodegenerative diseases and therefore offers numerous promising applications. 

Safe and fair?

- How can we make sure these expansions of data-intensive practices are safe and fair?

- Within this field, the thesis focuses on two key aspects of an AI production chain: the increased attention to the production and curation of referential datasets as ‘ground truths’ to train and evaluate algorithms, and the assembly of (modular) ‘algorithmic techniques’ using increasingly ‘off-the-shelf’ AI toolkits. Such data and algorithmic work are followed through the historical trajectories of voice AI in Europe; from the improvement of human-computer interaction (HCI) by the first European-funded projects and their focusing on understanding and reading emotion computationally – to the latest incursion of voice AI into bioinformatics and precision medicine through the development of  ‘vocal biomarkers’ that use the voice as a proxy to diagnose disease, that use speech recognition and LLM as well as speech emotion recognition, says Tanja Kanus. 

For the analysis of AI knowledge practices and its infrastructural aspects, Knaus conducted 20 interviews with industry professionals in Europe, analyzed technical documents, including computer science research articles that evaluate code, reviewed articles in the field, advertisement materials, and academic project websites, as well as the analysis of the most commonly used open-source benchmark datasets designed to evaluate algorithms. The key findings suggest that AI restructures and reconfigures how the body and mind are studied and consequently known. 

- So what have you accomplished?   

- I have contributed new insights into how Voice AI developed specifically in Europe, through major EU funding schemes and distinct researchers in experimental and computational psychology, who re-centered their research agenda to improve human-computer interaction (HCI). Their distinct scholarly engagement bridged psychology and HCI and provided insights for both disciplines, making these early developments an interesting case study that shows how voice AI developed differently in Europe, with heightened ethical awareness. 

- I also provided insights into how data-intensive voice AI applications operate in health care and bioinformatics today – how the industry is expanding the scope from these early HCI ambitions to a much more intricate and complex field, for example, entering the highly regulated clinical trial industry. I showed how such a push into health care could lead to the potential misuse of diagnostic technologies in secondary markets while harvesting sensitive patient data for training these models, with implications for AI policy, says Tanja Knaus. 

The trip to the UK 

For this type of research, stay at the Leverhulme Centre for the Future of Intelligence at the University of Cambridge in the UK was particularly fruitful and insightful, as the centre’s research projects address the challenges and opportunities of artificial intelligence, ensuring that AI is a force for good. The centre works on a wide range of projects, with a very interesting and diverse group of researchers from disciplines such as machine learning, philosophy, history, literary studies, engineering, media studies, and design. What was also interesting is that these research groups do keep close links with industry and policymakers, for a greater impact of their research, beyond academic merits.

- I was part of the research project “AI Narratives and Justice”, which investigates the cultural contexts shaping how AI is perceived and developed, and the consequences for diversity, cognitive justice and social justice. It is a program that brings together expertise from the humanities and social sciences, as well as computer science, to inform academia, industry, business, and government—for example, by collaborating with the UK government’s AI Security Institute (2025-2026), researching the risks of AI in critical infrastructure and societal systems - like financial markets, healthcare, and transportation. It was interesting to see how AI safety was very much interlinked with AI ethics to ensure ethical AI is developed, Tanja says.  

One of her contacts there is Maya Indira Ganesh, who recently published the book "Auto-Correct: The Fantasies and Failures of AI, Ethics, and the Driverless Car".

- This is an interesting read that I would highly recommend, in which Ganesh discusses attempts to automate ethical decision-making rather than remaining a human, social, or individualised practice. I would also highly recommend the very popular podcast “The Good Robot” by two feminist scholars who are also part of this research group: Dr. Elenor Drage and Dr. Kerry McINerney.

- What was your day-to-day life like?

- While there, I was attending the biweekly meetings, regular presentations by visiting scholars and other external researchers, attented CFI events, book launches and had regular lunch meetings with the program leader. During my time there, I kept thinking more about AI narratives and how to understand the narratives and metaphors that are used to describe AI functionalities – and how those are often inspired by very distinct human faculties that might be far from what AI is capable of, and sometimes misleading descriptors. I believe it is essential to consider how we talk, write, and promote AI, as these AI imaginaries do co-create the world we live in. The stay at CFI inspired my thinking and writing and fostered new, valuable connections with an interesting and important research group, says Tanja Knaus. 

By Rolf D. Svendsen

2025-12-17