Sarenne Wallbridge

Research Scientist

UK AISI | Human Influence

University of Edinburgh | ILCC & CSTR

About Me

I’m a research scientist on the Human Influence team at the UK AI Security Institute where I work to understand and address the socio-cultural impacts of AI on individuals and society at large. My main focus is the development of responsible audio AI; particularly assessing the capabilities of speech technologies and their impacts on human behaviour.

Before this, I was a research associate in the University of Edinburgh’s Institute for Language, Cognition and Computation. I completed my PhD in the Centre for Speech Technology Research at the University of Edinburgh where I worked with Catherine Lai and Peter Bell on information theoretic models of speech perception. I also worked on a number of start-up teams in healthtech, ML & NLP across the UK, Denmark, and Germany. Check out my resumé for more details (updated Nov 2025).

Alongside speech safety, my research explores the congitive mechanisms that underpin our ability to understand and produce language. During my PhD, I was particularly focused on how information encoded in the lexical and non-lexical channels of spoken communication affects perception. I applied these insights to better model human language comprehension as well as guide the development of efficient, perceptually-motivated representations of multi-channel communicative signals.

Translation: Spoken conversation is one of the most fundamental means we have for information transmission. How can we make use of information encoded in both [what words we say]_Lexical and [how we say them]_Non-lexical to better understand human communication and augment human-machine interactions?

In my spare time, I play, sing, and listen to music. I love climbing moutains and rocks (whichever are closer!) and teaching yoga when I can!

Interests

Speech Processing
Information Theory
Psycholinguistics
Machine Learning
Computational Linguistics
Natural Language Processing

Education

BSc (Hons) Cognitive Science, 2018

University of Edinburgh

News

06/2026: We released RealityTest, a benchmark for evaluating how AI systems respond when people probe their indentity in spoken and written interactions across 5 major languages. Our paper (and blog post) highlight the importance of understanding how people actually manage ambiguity about the identity of their partner in interactions for evaluating model behaviour.
03/2026: Sean will be presenting our paper at EACL in Morocco, go say hi :)
02/2026: I’ll be attending the International Association for Safe and Ethical AI'26 in Paris to talk about our work on how AI systems disclose their identity. Reach out if you’d like to arrange a chat!
12/2025: I’m organising the Young Female Researchers in Speech workshop again at Interspeech 2026, feel free to get in touch if you (or a student you know) are interested in applying!
12/2025: I’m also organising the Explainability for Compliance and Trust in Speech AI, a special session at Interspeech 2026, feel free to get in touch if you’d like to submit a paper!
09/2025: Our paper was accepted at ASRU, where I’ll also be chairing the special session on Responsible Speech and Audio Generative AI. Send me a message if you’d like to arrange a chat.
06/2025: I’ve joined the AI Security Institute in London as a research scientist on the Societal Impacts team!
05/2025: Our paper was accepted to Interspeech; blij om weer in NL te zijn!
01/2025: Our paper was accepted at ICASSP!
07/2024: I had the joy of supervising Sean Leishman’s MInf thesis; we turned his experiments into a long paper which was accepted at SemDial!
07/2024: New pre-print with UvA and ETH! This work presents a new paradigm for computing predictability during incremental language comprehension.
06/2024: Both of our papers were accepted at Interspeech–reach out if you’ll also be in Kos.
01/2024: Presenting a talk at the Speech, Music and Hearing group (TMH) in KTH Royal Institute of Technology, Stockholm.
10/2023: Our paper was accepted to EMNLP, shiok!
06/2023: Joined REST-CL–thanks for the invitation!
05/2023: Our paper was accepted at Interspeech, see you in Dublin!
04/2023: I’ll be on a research visit at the Dialogue Modelling Group of ILLC in Amsterdam.
03/2023: Presenting a talk at the CLS in Universiteit van Amsterdam
01/2023: Our paper was accepted at EACL 2023, pozdrav ako si u Dubrovniku!
01/2023: I’m running a lecture series An Introduction to Neural Networks on the Speech & Language Processing MSc this term.
12/2022: Presenting a talk at SIGMEDIA in Trinity College Dublin.
09/2022: Awarded Best Student Paper at Interspeech 2022!
05/2022: I’m organising UK Speech in Edinburgh this year (9/22). Come and say hi!