LLMs can predict psychological well-being from speech
A new study demonstrates that large language models (LLMs) can predict psychological well-being from spontaneous speech with a correlation of up to 0.8.
What happened?
Researchers have investigated the ability of large language models to predict Ryff Psychological Well-Being (PWB) scores from spontaneous speech. The study utilised voice recordings from 111 participants in the PsyVoiD database, testing several instruction-tuned LLMs, including variants of Llama-3, Mistral, and Gemma-2-9B. A domain-specific prompt was developed in collaboration with experts in clinical psychology and linguistics to optimise the analysis.
Key facts
| Publikationsdatum | 2026-05-11 |
|---|---|
| Antal deltagare | 111 |
| Högsta korrelation | 0.8 (Spearman) |
| Databas | PsyVoiD |
| Antal testade LLM | 12 |
”We investigate the use of Large Language Models (LLMs) for zero-shot prediction of Ryff Psychological Well-Being (PWB) scores from spontaneous speech.”
”Results show that LLMs can extract semantically meaningful cues from spontaneous speech, achieving Spearman correlations of up to 0.8 on 80% of the data.”
Why it matters
This research indicates technical potential for the early detection of changes in mental well-being based on speech analysis via LLMs. The ability to extract semantically meaningful signals from speech could offer new tools to complement traditional assessment methods. The results show that LLMs can function as a method for detecting linguistic patterns that correlate with psychological well-being.
Who is affected?
Researchers in psychology and linguistics are affected, as the study presents new methods for analysing speech and predicting mental well-being. AI model developers can benefit from insights into which LLMs perform best for this type of task. Future application developers in digital health may also be impacted, provided the technology's reliability and ethical aspects are further developed.
What else you should know
The study included analyses to characterise prediction variation and systematic biases, as well as keyword analyses to highlight which linguistic features drive the models' predictions. This contributes to a greater understanding of how LLMs make their decisions in this context.
Quick answers about this story
Vad har hänt?
När hände det?
Varför spelar det roll?
Vilka LLM testades?
The link opens in a new window and leads to the publisher's own site.
Källan har spårats automatiskt från utgivaren via Aheadlines signalkedja.
AI-verktyg i artikeln
Topics
Get similar news straight to your inbox
The reader's room
Send in a question or an addition. The newsroom reads everything before it's published and replies when relevant. No AI-generated text – just people.
Sign in to submit a comment or question.
Read the article through your role
- Decide whether this affects strategy over 6–12 months or is just noise.
- Discuss with leadership: do we own the right question or does ownership need to move?
- Ask: what risk are we taking by NOT acting on this this quarter?
Generated angle — not editorial analysis of "LLMs can predict psychological well-being from speech"