Social Simulations

We are working on creating an agent-based simulations to create an environment where we can test the impact of new policies or inspection strategies before they are deployed in the real world.

Inspectorates are vital to our society to ensure that companies, institutions and people operate safely and adhere to laws and regulations. Because they only have limited resources, inspectorates cannot inspect every company or institution. Consequently, they need to find strategies that effectively allocate their resources, such as the target selection. In addition, they need to ensure that their strategies are fair, robust, and actively help in creating better and safer environments.

Key Challanges in this domain

It is difficult for inspectorates to track the impact of their actions. Gathering insights into the behavioral dynamics of the inspectee population, such as network effects and inspection pressure, requires considerable effort. The same goes for tracking inspectees' decision-making and the practices they implement in an effort to increase their quality. Lastly, while inspection strategies are constantly re-evaluated, testing novel strategies in the real world is difficult if not impossible in many cases.

Key Questions

  • How we can optimize inspection strategies and balance the different objectives of an inspectorate?

  • How do inspectees influence each other?

  • How do inspectees influence each other?

  • What can we learn about non-compliant but not inspected organizations?

Solutions

We propose to build AI-driven tools that seamlessly integrate fairness and transparency into inspection planning. Rather than delivering a single, black-box recommendation, our system offers a range of balanced options that reflect different trade-offs between effectiveness, equity, and clarity. Users can explore these options through intuitive visual interfaces that highlight fairness metrics and provide clear explanations for why certain cases are prioritized, empowering inspectorates to make informed and trustworthy decisions.

I have always been interested in understanding dynamics of social systems, specifically the emergence of social phenomena. My previous experience lies within modelling multi-agent logic scenarios, so the step of modelling real scenarios poses an exciting challenge for me. I am glad the AI4Oversight lab gives me the opportunity to explore this topic, and to learn about and collaborate with the oversight domain.

Meet the researcher

Ira Adiprasito

Utrecht University

"Public safety, health, and equity are very important values to me. Utilising my dual background in psychology and artificial intelligence to further these values and improve social cohesion has long been an aspiration of mine. As inspections strive to ensure trust and accountability, improving these processes through our research allows me to put my background to use and contribute to society in a meaningful way."

Results

Inspectorate Use Cases

Description of use cases that have been executed within this work packages

Publications

Check out the publications related to Social Simulations