Skip to main content

Project RAAIT

Responsible Applied Artificial InTelligence (RAAIT) is a SPRONG research collaboration between the universities of applied science of Utrecht, Rotterdam, and Amsterdam. Goal of the program is to develop and validate methods for responsibly using and developing AI in practise.

Objectives​

The DSP has supported RAAIT, amongst others, in the following:

  • Writing a State of the Art section for a RAAK-PRO subsidy request. For a RAAK-PRO subsidy the DSP has supported in the writing of the State of the Art (SotA) section, including the creation of fitting figures. Topics within the SotA section included the development pipeline of Foundation Models, what control mechanisms on the FM's ultimate behaviour exists within each step of the development process, and the effect of AI model usage on the experience of agency and ownership of created content - specifically for Creative Media Professionals.
  • Conducting a feasibility study on the use of spatial data to fine-tune a state of the art Foundation Model. A research project between the HU and the Province of Utrecht, with support from SURF, this feasibility study looks at the use of destination plans ('bestemmingsplannen') for the fine-tuning of a Foundation Model (FM). Specifically, interest lies in the question whether the Dutch digital archive of destination plans (spanning over 12 years, being one of the most rich spatial-planning datasets worldwide) could be used to fine-tune a FM to reason over the spatial domain. The DSP is in charge of developing the model and documentation.

Contact​

See raait.nl for more information about RAAIT.

Contact Fabian (fabian.kok@hu.nl) for more information about the specific projects listed and the role the DSP plays within RAAIT.