AI, government and behavior

Artificial Intelligence (AI) is rapidly changing government organizations across the globe. AI is now used to support important tasks ranging from designing new data-based policies to detecting evidence of crimes in huge sets of data.

Big data © iStockphoto.com/Erikona

Balancing efficiency and public values

While AI-based technologies bring great promise to make governments more efficient, there is need to prevent bias and safeguard public values such as equity and fairness. To balance efficiency and public value we need an interdisciplinary approach from design oriented fields (data science, computer science), the social sciences (public administration, organizational science) and behavioral sciences to appreciate how AI systems will affect decision-making behavior in the complex setting of public organizations.

Joining forces

This SIG aims to foster closer collaboration between the focus areas of Human-centered Artificial Intelligence (HAI) and Governing the Digital Society (GDS). In this joint SIG, scholars from computer science, behavioral science and public administration join forces to contribute to the trustworthy use of AI-powered technologies in public organizations, such as the police, ministerial departments and municipalities.

The main ambition of this joint SIG is creating a ‘two-way street’ for computer scientists, behavioral scientists, and social scientists to contribute to AI systems that will improve behavior and decision-making of government workers.

Research questions

This SIG aims to investigate a broad set of questions, such as:

  • Can transparent/explainable AI improve decision-making by government workers? How can we design such systems from a socio-technical perspective?
  • Can we prevent behavioral tendencies such as automation bias by proper design?
  • To what extent are government workers using an AI system held accountable for their decisions? How can we design systems that enable such accountability?
  • What organizational principles are needed to ensure trustworthy AI use by government workers? How can such principles be integrated in the design of such systems?
  • What kind of behavioral interventions (training, nudges, boosts) can be developed and implemented to improve governmental use of AI applications?

Interested in participating?

Contact one of the principal researchers of this SIG!