Zoek in de site...

AI@EDU Infrastructure

The AI@EDU infrastructure is a novel approach and methodology for large-scale data collection in schools. The AI@EDU infrastructure will allow us to plug artificial intelligence into existing ALTs that are used in schools. In this way we can develop and refine AI solutions within schools without disrupting the curriculum or day-to-day school activities. This constitutes a unique approach to realizing theEuropean ethics-by-design vision for trustworthy, transparent and accountable applications of AI in public domains

The AI@EDU infrastructure methodology proposes an unprecedented approach to implementing AIs into existing educational contexts. This will enable a transparent approach to developing AI in which all stakeholders (schools, learners, teachers, publishers and Edtech developers) are included in the design, refinement and evaluation phases. As the project is publicly funded, open dialogue around the development will not be restricted by proprietary issues. This not only supports the trustworthiness of the solution, but also supports the discussion around accountability and ethics of applying AI in educational contexts. Moreover, to ensure future application and extended development of the Hybrid Human-AI Regulation framework, the code will be open access making this infrastructure a shared public good.

In order develop different new additions to ALTs used in schools, such as HHAIR and W-ALT, the AI@EDU infrastructure will be developed. This infrastructure will allow us to connect new apps and algorithms to ALTs used in schools, enabling flexible implementation within schools.

The AI@EDU infrastructure and collaboration

  • Data exchanges include advanced research collaboration between schools, companies and researchers. Schools and companies provide data about how their learners and users engage with learning technologies in order to co-create, validate and implement new AI driven technologies.

We have used this approach successfully before in collaboration with LMS provider Moodle, tool companies Hypothesis, AlgabraKit, Edtech and adaptive learning technologies Snappet, Gynzy, FutureWizz (Scula and WTRS), Learn! and Bettermarks. For example, in work with over 60 primary schools and adaptive learning technology Snappet, we evaluated how technical innovation supported and interacted with pedagogical innovation in the classroom. We described how this  changed teachers' professional routines (new lesson schema’s,  different feedback allocation ect) and analyzed children learning based on the data from the company (Molenaar et al. 2018; Breakthrough project).

AI@EDU infrastructure: An example

Direct data exchanges have been established with adaptive learning technology Gynzy through the AI@EDU infrastructure (Molenaar et al. 2020). This allows us to develop a new personalized dashboard for learners (Iprogress/learning path app), test new algorithms on company data. It also allows schools to engage in research within the learning technologies they use on a daily basis. This AI@Edu infrastructure not only allows for large-scale data collection in schools, it also allows us to plug artificial intelligence into existing adaptive technologies that are used in schools. In this way we can develop and refine AI solutions within schools without disrupting the curriculum or day-to-day school activities. This constitutes a unique approach to realizing an ethics-by-design vision for trustworthy, transparent and accountable applications of AI in the educational domain.