Neurofactor
Transparency

AI statement

We use artificial intelligence as a tool, never as a replacement for scientific judgement. This statement explains where AI sits in our work, where it explicitly does not, and who remains responsible.

Last updated: 1 September 2026

1. Our starting point

AI accelerates work; it does not determine conclusions. Every interpretation, recommendation and publication by Neurofactor is reviewed and approved by a human before it leaves the building.

We are transparent about the role of AI, because trust in research depends on traceability.

2. Where we use AI

  • Signal processing and pattern recognition in EEG, eye tracking and behavioural data, using scientifically grounded methods validated by our researchers.
  • Structuring and summarising large volumes of text, such as interviews, open answers and literature.
  • Translating our own content into Dutch, English and German, always with human final editing.
  • Support for text, imagery and visualisations on this website and in internal documentation.
  • Quality control, such as detecting inconsistencies in datasets and reports.

3. Where we do not use AI

  • Generating research results, figures or quotes that do not come from actual measurements.
  • Making decisions about people, such as selecting, scoring or rejecting candidates.
  • Making claims about individual participants based on their brain data.
  • Sending advice or reports to clients autonomously, without human review.
  • Uploading identifiable participant data to public AI services.

4. Data and AI

Raw, identifiable research data is not sent to external AI models. Where we use AI services, we do so with pseudonymised or aggregated data, with vendors that do not use customer data to train their models.

Confidential client information falls under the same rule: no input into public models, only environments covered contractually.

5. Human responsibility

A Neurofactor researcher is and remains accountable for every conclusion. If AI suggests a pattern the measurement design cannot support, we take it out of the report.

When in doubt, the scientific principle applies: a cautious claim is better than an attractive one that does not hold.

6. Bias and limitations

AI models inherit the biases of their training data. We therefore always test AI-supported outcomes against the raw data, the measurement context and the judgement of several researchers.

Where AI plays a substantial role in an outcome, we say so in the report.

7. AI-assisted content on this website

Parts of the texts, translations and illustrations on this site were produced with AI support and then edited and checked by our people. Research results, cases and figures always come from real projects.

8. Questions

Want to know what role AI played in a specific project? Ask us at contact@neurofactor.nl. We explain our approach per project.