Complex Social Behaviour Through the Lens of AI: Notes on the Future of Decision Support

Strategic decision-making is becoming increasingly complex, so there is a need to update decision-makers’ analytical toolkits to better understand and manage this complexity – and to help them make rational and well-informed decisions. Efforts to radically rethink how we build and use decision-support tools by applying AI are already underway. However, not all current decision-support tools account for important factors linked to the social complexity of human decision-making that social scientists examine to understand and forecast the possible outcomes of decisions and flashpoints.

The Alan Turing Institute held a 2-day workshop exploring long-term fundamental research challenges for the next 5–10 years that will significantly advance the performance of decision-support tools. The workshop was sponsored by the US Department of Defense Office of the Under Secretary of Defense for Research and Engineering through Virginia Tech Applied Research Corporation and was carried out in collaboration with the University of Maryland’s Applied Research Laboratory for Intelligence and Security.

The workshop convened AI technical experts, computational behavioural scientists, anthropologists, political scientists, defence strategists and professional wargaming experts. They highlighted the following six long-term fundamental research challenges:

1. Epistemological and ontological assumptions underpinning how humans understand social reality.
2. Optimising data foundations for AI.
3. Causality between choices and events.
4. Human-machine interaction.
5. Adversaries’ AI-based decision support.
6. Governance for future decision-support tools.