Please note: The algorithm descriptions in English have been automatically translated. Errors may have been introduced in this process. For the original descriptions, go to the Dutch version of the Algorithm Register.
AI analysis tool (AI Sensemaking)
- Publication category
- Other algorithms
- Impact assessment
- DPIA
- Status
- In use
General information
Theme
Begin date
Contact information
Link to publication website
Responsible use
Goal and impact
To facilitate understanding, categorisation and analysis of large amounts of input from residents within participatory processes. Only the content of contributions will be shared, not personal data.
Considerations
Using the tool saves a lot of time, gives a better understanding of the data and the built-in transparency keeps people able to control the content of the analysis. Only the content of the input is shared, the personal data is not (unless a resident shares personal data in their input).
Human intervention
The application is built entirely according to the human-in-the-loop philosophy. Read more here: https://support.govocal.com/nl/articles/8316692-ai-analyse
Risk management
The user interface alerts the user to errors or hallucinations of the generative AI. The user interface is built around the human-in-the-loop concept, where the raw source data is displayed on the screen at all times while generating summary. This encourages the user to control the algorithm and give their own interpretation where desired. Summaries can be deleted and modified. Summary includes references to source ideas.
Data is minimised; no PII is structurally sent to 3rd parties.
Legal basis
General data protection regulation (AVG)
Impact assessment
Operations
Data
Ideas, contributions (comments), survey results
Technical design
The functionality that generates summaries uses a Large Language Model hosted by Microsoft Azure. The collected contributions are structurally forwarded to the model, and the model is prompted to generate a complete summary, with references to relevant contributions where applicable.
The auto-tagging functionality for clustering contributions uses a Large Language Model to detect and classify themes. Classical classification algorithms are used for detecting sentiment and language.
External provider
Link to code base
Similar algorithm descriptions
- Data analysis for preparation and analysis of demonstrations.Last change on 26th of September 2024, at 11:49 (CET) | Publication Standard 1.0
- Publication category
- Impactful algorithms
- Impact assessment
- Field not filled in.
- Status
- In use
- This algorithm helps Customs to select goods for control based on customs declarations and risks. It uses declaration data from companies and considers whether or not there are increased risks of money laundering and terrorist financing for cash and valuables entering or leaving the European Union.Last change on 9th of December 2024, at 14:05 (CET) | Publication Standard 1.0
- Publication category
- Impactful algorithms
- Impact assessment
- Field not filled in.
- Status
- In use
- The algorithm is used by about 50 municipalities and, based on read-in data and answers given by the applicant, determines whether the applicant is eligible for any of the benefits to be applied for.Last change on 4th of April 2024, at 8:47 (CET) | Publication Standard 1.0
- Publication category
- Impactful algorithms
- Impact assessment
- Field not filled in.
- Status
- In development
- The algorithm is used by about 50 municipalities and, based on read-in data and answers given by the applicant, determines whether the applicant is eligible for any of the benefits to be applied for.Last change on 6th of September 2024, at 10:51 (CET) | Publication Standard 1.0
- Publication category
- Impactful algorithms
- Impact assessment
- Field not filled in.
- Status
- In development
- The algorithm is used by about 50 municipalities and, based on read-in data and answers given by the applicant, determines whether the applicant is eligible for any of the benefits to be applied for.Last change on 15th of November 2024, at 8:20 (CET) | Publication Standard 1.0
- Publication category
- Impactful algorithms
- Impact assessment
- Field not filled in.
- Status
- In use