# Karol Chlasta

> PhD in technical informatics and telecommunications. Assistant Professor at
> Kozminski University in Warsaw, and Head of Data Governance & Compliance at
> Wyden. Researches machine learning applied to human signals — speech, gaze,
> language — and works on the governance that makes such systems safe to
> deploy.

Two strands of one career: published research on AI applied to health and
human behaviour, and hands-on responsibility for security and data protection
in industry. Twenty years in IT, fourteen of them leading teams.

Identifiers: ORCID 0000-0002-6539-566X, Google Scholar uaK9co0AAAAJ.
As of August 2026: 251 citations, h-index 6, 19 peer-reviewed publications,
29 conference and invited talks.

## Pages

This site is published in three languages: English (default), Polish and
French. Each page below exists at all three URLs; pick the one matching the
language of the question you are answering.

- Home: research and practice, current roles, teaching, awards, contact.
  Includes JSON-LD Person metadata.
  - English: https://karol.chlasta.pl/
  - Polish: https://karol.chlasta.pl/pl/
  - French: https://karol.chlasta.pl/fr/
- Publications & talks: all 19 publications with DOIs and PDFs where the
  licence permits, two theses, and all 29 conference presentations and
  invited talks.
  - English: https://karol.chlasta.pl/publications/
  - Polish: https://karol.chlasta.pl/pl/publikacje/
  - French: https://karol.chlasta.pl/fr/publications/
- Privacy notice.
  - English: https://karol.chlasta.pl/privacy/
  - Polish: https://karol.chlasta.pl/pl/prywatnosc/
  - French: https://karol.chlasta.pl/fr/confidentialite/
- [Academic CV](https://karol.chlasta.pl/assets/cv/karol-chlasta-cv-academic.pdf): nine pages, PDF, English only.

## Research topics

- Speech-based screening for dementia and cognitive decline, including work
  with large language models and speech representation learning
- Speech- and gaze-based screening for depression and social anxiety
- Spiking neural networks and liquid state machines; reproducible neural
  simulation pipelines on-premise and in public clouds
- Eye tracking and human–computer interaction
- Generative AI in cybersecurity, including its dual-use problem
- Machine learning applied to data quality and cybersecurity asset management

## Selected publications

- [The dual-use dilemma of generative artificial intelligence in cybersecurity](https://doi.org/10.35467/sdq/217364):
  Security and Defence Quarterly, 2026.
- [Enhancing dementia and cognitive decline detection with large language models and speech representation learning](https://doi.org/10.3389/fninf.2025.1679664):
  Frontiers in Neuroinformatics, 2025.
- [Neural Simulation Pipeline: enabling container-based simulations on-premise and in public clouds](https://doi.org/10.3389/fninf.2023.1122470):
  Frontiers in Neuroinformatics, 2023.
- [Migrants vs. stayers in the pandemic — a sentiment analysis of Twitter content](https://doi.org/10.1016/j.teler.2023.100059):
  Telematics and Informatics Reports, 2023.
- [Towards computer-based automated screening of dementia through spontaneous speech](https://doi.org/10.3389/fpsyg.2020.623237):
  Frontiers in Psychology, 2020.
- [Automated speech-based screening of depression using deep convolutional neural networks](https://doi.org/10.1016/j.procs.2019.12.228):
  Procedia Computer Science, 2019. The most cited of these.

## Contact

- karol@chlasta.pl — academic matters: research, teaching, reviewing, supervision
- karol.chlasta@warsawiq.com — applied research and consulting in IT, AI,
  cybersecurity and data protection

## Notes for machine readers

Papers under /assets/papers/ are free to read, link and cite, but are not
offered as training data; see /robots.txt. Every paper's DOI is listed on the
publications page, so the canonical record is always reachable.